Faculty 16 "Mathematics, Informatics and Statistics": Research Report 1998 -- 2003
11 February 2004edited by François Bry, dean for research
Die Mathematik ist eines der wichtigsten Grundlagenfächer der LMU und zwar nicht nur als grundständiges Fach in der Lehre, bei der Lehrerausbildung und als Dienstleister in sämtlichen naturwissenschaftlichen Fächern, sondern sie ist vor allem auch in der Forschung integraler Bestandteil von zukunftsträchtigen Fächern wie der Biotechnologie, der Genforschung und den Nanowissenschaften.
(A. Heldrich, Rektor der LMU, August 2001)
Der vorliegende Bericht gibt einen Überblick über die Forschungsaktivitäten am mathematischen Institut von Mitte 1999 bis Ende 2003. Er ist gegliedert in Berichte der einzelnen Arbeitsgruppen, sowie der beiden im Berichtszeitraum am mathematischen Institut angesiedelten interdisziplinären Graduiertenkollegien: ``Mathematik im Bereich ihrer Wechselwirkung mit der Physik'' und ``Logik in der Informatik''.
In diesen Zeitraum fällt eine wesentliche personelle Erneuerung des Instituts, die noch mehrere Jahre andauern wird. Ein Lehrstuhl für Angewandte Mathematik, insbesondere Numerische Mathematik, war seit der Emeritierung von Professor Dr. G. Hämmerlin am 1. April 1996 vakant. Er wurde erst im Sommer 2003 durch die Berufung von Professor L. Erdös Ph. D. wieder besetzt. Ein weiterer Lehrstuhl für Angewandte Mathematik, insbesondere Analysis und Mathematische Physik, war seit der Emeritierung von Professor Dr. E. Wienholtz zum 1. Oktober 1997 vakant; er wurde im Sommer 2000 durch die Berufung von Professor Dr. H. Siedentop wieder besetzt. Seit dem 1. April 2000 war der ehemalige Lehrstuhl von Professor Dr. H. Kellerer, der in einen Lehrstuhl für Topologie umgewidmet wurde, vakant. Er wurde im April 2003 durch die Berufung von Professor B. Leeb Ph. D. neu besetzt. Eine C3-Professur wurde mit Professor Dr. K. Cieliebak besetzt. Seit dem 1. April 2002 ist der Lehrstuhl für Stochastik von Professor Dr. P. Gänßler vakant; der Lehrstuhl für Finanz- und Versicherungsmathematik ist durch Wegberufung von Professor Dr. M. Schweizer an die ETH Zürich seit dem 1. Oktober 2003 ebenfalls vakant. Das Berufungsverfahren für beide Lehrstühle ist im Gange. Diese Vakanzen hatten und haben, trotz Lehrstuhlvertretungen, starke Auswirkungen in den betroffenen Arbeitsgruppen, und darüber hinaus im ganzen Institut.
Grundlage für die Neuberufungen ist der Strukturplan des Mathematischen Instituts von 1999, der einen Wechsel von einer lehrstuhlorientierten Struktur zu einer Struktur mit fünf größeren wissenschaftlichen Arbeitsbereichen sowie einem Lehrstuhl für Didaktik der Mathematik vorsieht. Außerdem sah der Plan die Einrichtung eines neuen Studiengangs für Wirtschaftsmathematik und Aktuarwissenschaft sowie eines internationalen Master Studiengangs vor; beides ist inzwischen realisiert. In dem vorliegenden Bericht erscheinen nun die neuen Arbeitsbereiche Analysis und Numerik, Differentialgeometrie und Topologie, und Stochastik und Finanzmathematik. Die Schaffung des Arbeitsbereichs für Algebraische und Arithmetische Geometrie (bisher einerseits Algebra, und andererseits Komplexe Analysis und algebraische Geometrie) steht noch aus.
Im November 1999 und nochmals im Frühjahr 2003 wurde das Mathematische Institut evaluiert. Für die Mitarbeit in der Kommission konnten führende Mathematiker von auswärtigen Universitäten und Max-Planck-Instituten gewonnen werden. Unter den Gutachtern befanden sich drei Leibniz-Preisträger der DFG sowie ein Inhaber der Fields-Medaille. Die Evaluatoren befürworteten in ihrem Bericht vom Februar 2000 nachdrücklich den geplanten Strukturwandel sowie die Einrichtung der neuen Studiengänge. In ihrem erneueten Bericht vom Mai 2003 schreiben sie:
Seit der letzten Evaluation hat das Mathematische Institut eine äußerst erfreuliche Entwicklung genommen. Das wissenschaftliche Profil ist durch hochkarätige Berufungen (Siedentop, Schweizer, Leeb, Erdös, Cieliebak) wesentlich gestärkt worden. Damit verbunden war ein starker Anstieg der eingeworbenen Drittmittel. In einigen Gebieten der Mathematik befindet sich das Institut jetzt in der Spitzengruppe in Deutschland.
Wissenschaftler mit Dauerstellen: Prof. Dr. B. PAREIGIS, Prof. Dr. H.-J. SCHNEIDER, Prof. Dr. W. ZIMMERMANN, Prof. Dr. H. ZÖSCHINGER, Priv.-Doz. Dr. G. HAUGER (1#1 7. 7. 2001), Dr. E. SCHÖRNER (Seit 1. 10. 2002)
Forschungsschwerpunkte sind:
Anwendungsgebiete solcher sequenziell dynamischen Systeme sind u. a. aerodynamische Modelle (stetig oder differenzierbar), biologische Prozesse und Zellwachstum (diskret), Straßenverkehr in Städten (hybrid), Telefonsysteme, Paketverkehr in Datenleitungen, verteilte Rechnersysteme, und der Kapitalfluß auf den Weltmärkten.
Es besteht Bedarf nach einer mathematischen Grundlegung, mit der viele Fragen überhaupt erst zugänglich gemacht und systematisch behandelt werden können. So sind inzwischen die zulässigen Abbildungen, gewisse Kolimites von SDS und mit ihnen der Begriff der Simulation genauer untersucht worden. Diese Untersuchungen dienen der Konstruktion effizienter und optimaler SDS.
Auslander-Reiten-Folgen dienen der Konstruktion unzerlegbarer Moduln. Sehr interessant ist die Frage nach der Existenz solcher Folgen über Ringen, die keine Artin-Algebren sind.
at roots of unity,
Commun. Algebra 28 (2000), 1029-1044.
-Hopf algebras), in: ,,New
trends in Hopf algebra theory``, Hrsg.: N. Andruskiewitsch,
W. Ferrer-Santos, H.-J. Schneider, Tagungsband zu
,,Colloquium on quantum groups and Hopf algebras`` (Córdoba,
Argentinien, 1999), Contemp. Math. 267 (2000),
Am. Math. Soc., Providence.Wissenschaftler mit Dauerstellen: Prof. L. ERDÖS (Seit 1. 7. 2003), Ph.D., Prof. Dr. H. SIEDENTOP (Seit 1. 6. 2000) , Prof. Dr. H. KALF, Prof. Dr. W. RICHERT, Prof. Dr. A. SACHS, Prof. Dr. H. STEINLEIN, Priv.-Doz. Dr. E. SCHÄFER, Dr. J. SCHMALZING (Seit 1. 10. 2003, Leiter des Rechenzentrums), Dr. E. JÖRN (Bis 30. 9. 2003, Leiter des Rechenzentrums) , Dr. W. SPANN (Mitarbeiter des Rechenzentrums.)
Randwertprobleme für subordinierte Dirichlet und Neumann-Probleme wurden auf den Fall nicht glatter (fraktaler) Ränder übertragen.
Anwendungen dieser Sätze über atomare Zerlegungen von Funktionenräumen findet man in einer umfangreichen Behandlung aller Grenzfälle der 'Spuren von Funktionenräumen'. Die Kenntnis der genauen Spur eines Funktionenraumes spielt eine Schlüsselrolle in der Behandlung von partieller Differentialgleichungen.
Im Mittelpunkt zweier weiteren Arbeiten stand die systematische
Untersuchung der 
-Varianten von Funktionenräumen. Diese werden
benötigt um Pseudodifferentialoperatoren zu behandeln, die
-sub-Markoffsche Halbgruppen erzeugen, falls das
Symbol des Pseudodifferentialoperators mit einer stetig negativ
definiten Funktion vergleichbar ist.
. Enhanced binding,
Self-energy of an electron coupled to a photon field
)-Abbildungen. ist die adäquate Lie-Gruppe für periodische
Probleme mit räumlicher Symmetrie. Ziel der Untersuchungen
sind konkrete Formeln und Berechnungsverfahren für
verschiedene Typen von Verzweigungen, insbesondere in Verbindung
mit der Multiplikativitätseigenschaft des äquivarianten
Abbildungsgrades.
for any time 
.
dimensions. We prove this claim in the scaling
limit.-Equivariant Degree with
Applications to Symmetric Bifurcation Problems. Nonlinear
Anal. 47 (2001), 1617-1628
, Z. Anal.
Anwendungen 18 (1999), 875-893.
-sub-Markovian semigroups,
and applications to pseudo-differential operators with negative
definite symbols, Forum Math. 13 (2001), 59-90.
-Bessel potential spaces,
Dissertationes Math. 393 (2001), 1-63. and applications to N-body quantum Systems,
Lett. Math. Phys. 61 2002, 75-84.
from the nucleus, Math.
Z. 236 (2001), 787-796.-equivariant degree with
applications to symmetric bifurcation problems: the case of one
free parameter, Topol. Methods Nonlinear Anal. 20
(2002), 335-374Wissenschaftler mit Dauerstellen: Prof. D. KOTSCHICK, D. Phil., Prof. B. LEEB (Seit 1. 4. 2003') , Ph.D., Prof. Dr. K. CIELIEBAK (Seit 1. 10. 2001)
Schliesslich wurden Paare von komplementären Lagrangeschen Blätterungen auf (fast-) symplektischen Mannigfaltigkeiten untersucht.
-algebras, bordism-algebras (in cooperation
with T. Schick (Göttingen)) are studied. It is also
intended to establish new existence results using methods related
to equivariant bordism theory.
Structures and Scalar
Curvature Estimates, Ann. Global
Anal. Geom. 20 (2001), 301-324.Wissenschaftler mit Dauerstellen: Prof. Dr. O. FORSTER, Prof. Dr. H.-W. SCHUSTER, Prof. Dr. K. WOLFFHARDT (Bis 31. 3. 2003) , Prof. Dr. G. KRAUS, Prof. Dr. M. SCHOTTENLOHER
Wissenschaftler mit Dauerstellen: Prof. Dr. H. SCHWICHTENBERG, Prof. Dr. H.-D. DONDER, Prof. Dr. H. OSSWALD, Prof. Dr. W. BUCHHOLZ
mit
A(x,y) quantorenfrei in einen konstruktiven
Beweis derselben Formel umgeformt werden kann. Dieses theoretische
Resultat ist jedoch ohne Verfeinerungen für Anwendungen
unbrauchbar; derartige Verfeinerungen sind ein Gegenstand unserer
Untersuchungen.
-CA durch M. Rathjen und T.
Arai) hat einen extrem hohen Grad an Kompliziertheit und
mathematischer Tiefe erreicht, so dass weitere Untersuchungen
erforderlich sind, um zu einem besseren Verständnis der
betreffenden Ergebnisse und Beweise zu kommen. Durch Aufdeckung von
Zusammenhängen zwischen bestimmten Operationen auf unendlichen
bzw. endlichen Herleitungen wird versucht, einen Beitrag zu diesem
Ziel zu leisten. -- Die finite Darstellung unendlicher Herleitungen
war auch Ausgangspunkt für Untersuchungen zur Komplexität
von durch höherstufige Rekursionsgleichungen definierten
Funktionalen. Dabei ergaben sich u.a. Zusammenhänge mit Themen
der theoretische Informatik (z.B. Co-Rekursion und -Induktion,
Produktivität von Strömen), denen weiter nachgegangen
werden soll.In meiner Arbeit wende ich diese Einsichten an auf die Maßtheorie (es entstehen neue Produkträume) und auf die unendlich dimensionale stochastische Analysis.
At the moment I apply this result to spaces of continuous functions, endowed with Gaussian measure, and to spaces of functions that are right continuous and have left hand limits, endowed with Poisson measure (Malliavin Calculus).
, Journal of Symbolic Logic
68 (2003) 713-739.Wissenschaftler mit Dauerstellen: Prof. Dr. P. GÄNSSLER (Bis 31. 3. 2002) , Prof. Dr. H. KELLERER (Bis 31. 3. 2000) , Prof. Dr. M. SCHWEIZER (Vom 6. 9. 2001 bis 30. 9. 2003) , Prof. Dr. D. DÜRR, Prof. Dr. H.-O. GEORGII, Prof. Dr. U. OPPEL, apl. Prof. Dr. H. PRUSCHA
. Durch eine gewichtete, gleitende Mittelwertbildung
(mittels eines Fensters mit datengesteuerter Breite) entlang der
bivariaten (x,y)-Stichprobe gewinnt man einen
nichtparametrischen Kurvenschätzer für 
. Von besonderem Interesse
sind die Stellen, an welchen die Funktion maximale Werte aufweist
(modes). Auf der Grundlage einer Stichprobenfolge ist
für den mode eine geeignete Schätzerfolge zu
finden und ein Konfidenzintervall zu konstruieren. Hierfür
benötigt man u.a.\ bootstrap Verfahren und die Theorie der
allgemeinen empirischen Prozesse.Wissenschaftler mit Dauerstellen: Prof. Dr. R. FRITSCH, Dr. G. STUDENY
Im Zusammenhang damit wird der Geometrie-Unterricht in der 8. Jahrgangsstufe von Gymnasien und Realschulen untersucht. Es geht um die unterrichtliche Behandlung der Sätze im Umfeld des Umfangswinkelsatzes, die Klassifikation der konvexen Vierecke im Haus der Viereche und die Grundlegung der räumlichen Geometrie, für die es noch keinen systematischen Aufbau gibt, der über die Behandlung in Euklids Elementen hinausgeht.
, Der
Mathematikunterricht 48/1 (2002), 57-63
-Quadrate als Beispiel für einen Vektorraum,
PM Praxis der Mathematik in der Schule - Sekundarstufen 1 und
2 44 (2002), 129-133, Differentsial'naya Geometriya Mnogoobrazij
Figur 34 (2003), 144-148FRITSCH R.: EUR 6.500 als Spenden aus Anlass des Tages der Mathematik 2003 erhalten von der BMW Group, der Münchner Rück, der Siemens AG und der Stadtsparkasse München
Auslauffinanzierung:
1. 4. 2002 bis 31. 3. 2003
Ab 1.4.2000 Zusatzförderung (siehe unten, Punkt 5)
Prof. Dr. Martin Schottenloher
angegeben):
Prof. Dr. Jürgen Batt
Prof. Dr. Detlef Dürr
Prof. Dr. Hans-Otto Georgii
Prof. Dr. Hubert Kalf
Prof. Dieter Kotschick, D. Phil.
Dr. Dieter Maison (MPI für Physik, München)
Prof. Dr. Bodo Pareigis
Prof. Dr. Hans-Jürgen Schneider
Prof. Dr. Martin Schottenloher
Prof. Dr. Herbert Spohn (TU München)
Prof. Dr. Heinrich Steinlein
Prof. Dr. Stefan Theisen (Sektion Physik, LMU
München)
Prof. Dr. Julius Wess (Sektion Physik, LMU München
und MPI für Physik)
Mit dem Antrag auf Einrichtung des Graduiertenkollegs
``Mathematik im Bereich ihrer Wechselwirkung mit der Physik'' wurde
beabsichtigt, der Entwicklung eines in neuerer Zeit wieder
stärkeren Zusammenwirkens von Mathematik und Physik Rechnung
zu tragen und den Trend der gegenseitigen Förderung der beiden
Wissenschaften zu verfolgen und diese weiter zu vertiefen. Es wurde
dabei aufgebaut auf bereits bestehende Kontakte zwischen
Professoren der Mathematik und Physik und auf ihre bisher
geleistete Zusammenarbeit. Diese interdisziplinäre Kooperation
ist im Rahmen des mittlerweile seit sechs Jahren bestehenden
Graduiertenkollegs weitergeführt und in eine neue
Organisationsform gebracht worden. Vom Angebot der Forschungsthemen
wie der Ausbildungsveranstaltungen her soll das Graduiertenkolleg
das Bewußtsein für die vielfältigen Verbindungen
und Wechselwirkungen zwischen Mathematik und Physik schärfen.
Schließlich wird in dem Graduiertenkolleg eine Vertiefung der
Zusammenarbeit der beteiligten Mathematiker und Physiker
angestrebt.
Den Graduierten (also Stipendiaten und assoziierten Kollegiaten)
soll im Graduiertenkolleg eine über das spezielle
Forschungsprojekt hinausgehende breitere Ausbildung vermittelt
werden. Besondere Formen des Erfahrungs- und Wissensaustausches
werden angeboten, die bewirken sollen, daß der Graduierte
zusätzliche Anregungen erhält. Durch das Forschungs- und
Ausbildungsprogramm des Graduiertenkollegs wird auch angestrebt,
den Doktoranden des Kollegs eine über die übliche
Betreuung durch den Doktorvater hinausgehende weitere Beratung und
Förderung durch die anderen beteiligten Hochschullehrer oder
auch durch Gastwissenschaftler zu geben. Schließlich soll
diese zusätzliche Förderung auch eine Straffung und eine
intensive Durchführung des Promotionsstudiums bewirken.
Das gemeinsame Forschungsprogramm ist dem Thema ,,Mathematik im Bereich ihrer Wechselwirkung mit der Physik`` gewidmet. Ziel des wissenschaftlichen Programms ist die Erforschung und Anwendung mathematischer Strukturen, die bei physikalischen Problemen zum Tragen kommen: Für das Graduiertenkolleg relevante Beispiele sind die Differentialgeometrie in der Allgemeinen Relativitätstheorie und der Quantenfeldtheorie, die Operatortheorie in der Quantenmechanik, die Stochastik in der Statistischen Physik, die Hopf-Algebren zur Formulierung von allgemeinen Symmetrien und Differentialgleichungen etwa bei der Behandlung von Bewegungsgleichungen der Physik. Ebenso gehört zum Forschungsprogramm des Graduiertenkollegs die Untersuchung der Anwendungen von Ideen und Prinzipien der Physik in der Mathematik. Solche Anregungen haben in den letzten 20 Jahren immer wieder zu interessanten mathematischen Ergebnissen geführt.
Das angestrebte Forschungsprogramm wird in den folgenden vier Schwerpunkten realisiert:
A: Differentialgleichungen und Funktionalanalysis
Ziel dieses Forschungsschwerpunktes ist die Weiterentwicklung
abstrakter Methoden der linearen und nichtlinearen
Funktionalanalysis und die Anwendung funktionalanalytischer
Methoden auf Differentialgleichungen und -operatoren der
Mathematischen Physik in den folgenden Arbeitsrichtungen:
1) Theorie der Grundgleichungen der Astrophysik (Stellardynamik), der Plasmaphysik (und der Relativitätstheorie) (am Lehrstuhl Batt). Seit 20 Jahren stehen im Vordergrund der Untersuchungen Fragen der Existenz und Eindeutigkeit klassischer und schwacher Lösungen des Vlasov-Poisson-Systems und des Vlasov-Maxwell-Systems partieller Differentialgleichungen, die die Evolution gravitierender Materie und von Plasmen (mit mehreren Komponenten geladener Teilchen) beschreiben, und von deren qualitativem Verhalten, ferner Fragen der Existenz stationärer Lösungen und deren Stabilität. Nach dem Abschluß des klassischem Existenztheorie durch K. Pfaffelmoser (München) gewinnen neuerdings singuläre Lösungen und ihr Bezug zu verwandten Systemen (wie dem Euler-Poisson-System) großes Interesse (hydrodynamischer Limes). Die behandelten Vlasov-Einstein-Gleichungen stellen einen Bezug her zu Differentialgeometrie (Forschungsschwerpunkt B), numerische Verfahren (Partikel-Methoden) einen solchen zur Wahrscheinlichkeitstheorie und Statistischen Physik (Forschungsschwerpunkt D). Die mathematischen Hilfsmittel entstammen u.a. der Nichtlinearen Funktionanalysis bzw. der Theorie der semielliptischen Gleichungen (die zur Arbeitsrichtung von Steinlein und Kalf desselben Forschungsschwerpunktes A gehören). In Zukunft sollen verstärkt die Hamiltonschen und die Symmetriestrukturen der Gleichungen untersucht werden. Bei diesen Fragestellungen ist eine Kooperation mit Prof. J. Scheurle (TU München) geplant.
2) Spektraltheorie von elliptischen Differentialoperatoren
(Schrödingeroperatoren) im Zusammenhang mit der
Quantenmechanik, quasilineare elliptische Gleichungen und
hyperbolische Systeme von partiellen Differentialgleichungen (bei
Kalf): Neuere Resultate betreffen Aussagen über das Spektrum
von Schrödingeroperatoren mit rotationssymmetrischem Potential
und Aussagen über Lücken im wesentlichen Spektrum (siehe
auch die Angaben von Kalf im Anhang 1). Herr Kalf verfolgt seit
Jahren die Weiterentwicklung der Arbeitsrichtung 1); semilineare
Gleichungen spielen u.a. bei der Konstruktion von stationären
Modellen des Vlasov-Poisson- und des Vlasov-Maxwell-Systems die
entscheidende Rolle. Die bestehende Zusammenarbeit auf diesem
Gebiet wird in das Graduiertenkolleg integriert.
3) Nichtlineare Funktionalanalysis und diskrete dynamische
Systeme (bei Steinlein). Im Bereich der nichtlinearen
Funktionalanalysis konzentrieren sich die Untersuchungen auf die
Theorie des äquivarianten Abbildungsgrades, d.h. eines
Abbildungsgrades 
für äquivariante Abbildungen f
zwischen G-Räumen (G eine kompakte Lie-Gruppe).
Interessante Fragestellungen sind Beziehungen zwischen
verschiedenen Abbildungsgradtheorien, konkrete Berechnungen
für spezielle Gruppen G und geeignete orthogonale
Darstellungen sowie Anwendungen insbesondere auf äquivariante
Multiparameterverzweigungsprobleme.
B: Differentialgeometrie, Komplexe Analysis und
Stringtheorie
Die Wechselwirkung zwischen Mathematik und Physik ist in der Differentialgeometrie (unter Einbeziehung von Eichtheorie und Differentialtopologie) besonders deutlich. Die Beziehung ist so eng, daß viele der benutzten Begriffe eine eindeutige Zuordnung zu einer der beiden Wissenschaften nicht zulassen. Viele der heute verwendeten Konzepte wurden zunächst in der Physik entwickelt, anschließend mathematisch eingehend untersucht und verallgemeinert und fanden später in ganz neuem Zusammenhang wieder Eingang in die Physik, oder sie führten zu unerwarteten Einsichten in mathematische Theorien. Diese Entwicklung hält unvermindert an.
Eines der eindrucksvollsten Beispiele dafür sind die Anwendungen der Eichtheorie in der 4-dimensionalen Topologie. Diese begannen 1982 mit den Arbeiten von Donaldson, der zeigte, daß die Lösungsräume der nicht-linearen Yang-Mills Gleichungen, d. h. die klassischen Grundzustände der Yang-Mills Theorie, interessante Information über die Differentialtopologie der zugrunde liegenden Mannigfaltigkeit enthalten. So findet man zum Beispiel exotische differenzierbare Strukturen auf dem 4-dimensionalen euklidischen Raum. Seit 1994 weiß man, daß solche Information in konzentrierter Form auch in den sogenannten Seiberg-Witten Gleichungen enthalten ist, die zu einer zu Yang-Mills dualen Quantenfeldtheorie gehören. Die mathematischen Anwendungen dieser und verwandter Ideen werden im Graduiertenkolleg intensiv untersucht.
Die klassische Allgemeine Relativitätstheorie stellt einen der Höhepunkte der oben beschriebenen Wechselwirkungen dar. Derzeit besteht großes Interesse an der Untersuchung ,,teilchenartiger Lösungen`` der Einstein-Gleichungen (selbst-gravitierende Solitonen, schwarze Löcher). Während für Theorien mit Abelschen Eichfeldern (z.B. Einstein-Maxwell-Theorie) die Lösungsmannigfaltigkeiten mit analytischen Methoden bereits sehr genau untersucht worden sind, liegen für Theorien mit nicht-Abelschen Eichfeldern nur Teilergebnisse vor. Das Spektrum der angewandten mathematischen Methoden reicht dabei von der aus der Theorie der Solitonen bekannten Inversen Streumethode (hier ergeben sich Bezüge zu den Forschungsschwerpunktem A und D) bis zu rein numerischen Lösungsmethoden. Wichtige offene Fragen sind mit Schlagworten wie ,,No Hair Conjecture`` und ,,Cosmic Censorship`` angesprochen.
Ein Ideenbereich, in dem die oben angesprochene Wechselwirkung ebenfalls von besonderem Interesse ist, ist die Konforme Feldtheorie, die außer zur Differentialgeometrie Beziehungen zu vielen anderen aktuellen Bereichen der Mathematik und der Physik aufweist, u. a. zur Statistischen Mechanik (und damit zum Schwerpunkt D), zur Theorie der Quantengruppen (und damit zum Schwerpunkt C), zur Knotentheorie, zur Stringtheorie und zur Komplexen Analysis (mit ihren Riemannschen Flächen, Modulräumen etc.). Die Komplexe Analysis hat in den letzten Jahren in verschiedenen Bereichen der Physik an Bedeutung gewonnen. Das trifft besonders für die Stringtheorie zu, in der Modulräume ein wesentliches Werkzeug geworden sind und in der Kählermannigfaltigkeiten und torische Varietäten eine große Rolle spielen.
Ein weiteres Arbeitsgebiet innerhalb des Schwerpunktes B ist die bereits erwähnte Stringtheorie, bei der zweidimensionale Konforme Feldtheorien eine zentrale Rolle spielen. In einer alternativen Formulierung der Stringtheorie stehen die algebro-geometrischen Aspekte im Vordergrund. Zum Beispiel sind bei der Kompaktifizierung von Raumzeiten in der Stringtheorie die Calabi-Yau-Mannigfaltigkeiten von besondere Bedeutung. Dieser Problembereich hat in den vergangenen Jahren zu einem regen Austausch von Ideen und gegenseitigen Stimulationen von Mathematikern und Physikern geführt (Stichwort: Spiegelsymmetrie) mit bemerkenswerten Resultaten auch für die Algebraische Geometrie. In jüngster Zeit hat man, stimuliert durch die Arbeiten von Seiberg und Witten, Dualitätssymmetrien in der Stringtheorie gefunden, die es erlauben, bisher als verschieden betrachtete Theorien als nichtperturbativ äquivalente Formulierungen ein und derselben Theorie zu betrachten. Bei diesen Identifizierungen spielen die sogenannten D-branes eine zentrale Rolle.
C: Quantengruppen und Nichtkommutative Algebra
Seit den Arbeiten von Drinfel'd über Quantengruppen (1987), von Connes über nichtkommutative Geometrie (1985), von Atiyah und Witten zur Quantenfeldtheorie (1988) und von Jones über Knoten-Invarianten (1985) hat sich ein großes fachübergreifendes Arbeitsgebiet mit den mathematischen Teilgebieten der Kategorientheorie, der Knotentheorie, der Hopf-Algebren (Quantengruppen) und den physikalischen Teilgebieten der Statistischen Mechanik, der Quantenfeldtheorie, sowie der Quantenmechanik etabliert.
Im Rahmen des Graduiertenkollegs wird innerhalb des Schwerpunktes weiterhin an der Erforschung der algebraischen Struktur von Quantengruppen, insbesondere der halbeinfachen und der punktierten Quantengruppen, der Aufdeckung der Zusammenhänge zwischen ihrem Auftreten bei Quasisymmetrie-Bedingungen, z.B. in monoidalen Kategorien, bei den Darstellungen der Zopf-Gruppen oder in der Quanten-Yang-Baxter-Gleichung und ihrem Auftreten als Symmetriegruppen von nichtkommutativen Räumen (Quantenräumen) gearbeitet. Dabei sollen auch physikalisch interessante Eigenschaften untersucht und verstanden werden. Ebenfalls dazu gehört die Untersuchung der Differentialrechnung auf Quantenräumen und ihre Einbettung in Quantengruppen. Es sollen insbesondere drei Bereiche schwerpunktmäßig untersucht werden:
1) Der Zusammenhang zwischen den Darstellungen von Zopf-Gruppen und den dabei auftretenden Knoten-Invarianten, den Quasi-Symmetrien in monoidalen Kategorien, den Lösungen der Quanten-Yang-Baxter-Gleichung und den quasitriangulären Hopf-Algebren (als Symmetriegruppen). Als Komplement hierzu sind in der Physik die Statistische Mechanik, die Theorie der Anyonen und niederdimensionale topologische Quantenfeldtheorien zu untersuchen. In dieses Gebiet gehören u.a. Supersymmetrie, Quantenkommutativität, Vektorsymmetrie, R-Matrix, Band-Graphen und Tannaka-Krein-Dualität.
2) Die aus der (algebraischen) Geometrie bekannte Konstruktion der Symmetriegruppe eines Raumes ist auch in der nichtkommutativen Geometrie möglich und führt zu Quantengruppen. Es soll die algebraische Struktur der Quantengruppen und ihrer homogenen Räume, insbesondere der Hauptfaserbündel mit einer Quantengruppe als Strukturgruppe und der verschränkten Produkte von Quantengruppen untersucht werden. Hierzu gehören die Untersuchung der Quotientenbildung, der Quanten-Deformationen, der Hopfalgebren in der verzopften Kategorie der Yetter-Drinfel'd Moduln über einer abelschen oder nicht-abelschen Gruppe, und der Verallgemeinerungen von Quantengruppen, z.B. Quasi-Hopf-Algebren (im Sinne von Drinfel'd). Für die physikalischen Anwendungen sind die möglichen Deformationen und die Darstellungen der physikalisch relevanten Gruppen besonders interessant.
3) Es bedingen sich gegenseitig die Operation der Quantengruppen auf einem Quantenraum und eine Differentialrechnung auf einem Quantenraum. Hier haben erste Untersuchungen stattgefunden, um die von Manin, Wess, Woronowicz, Zumino und anderen gefundenen Ansätze algebraisch zu verarbeiten und durch gewisse universelle Moduln zu erfassen.
Seit Beginn des Graduiertenkollegs hat sich die fächerübergreifende Zusammenarbeit in dem Forschungsschwerpunkt erfreulich positiv entwickelt. Im gemeinsamen Seminar Pareigis/Wess sowie bei vielen Veranstaltungen mit Gästen kam es zu produktiven gegenseitigen Anregungen.
D: Wahrscheinlichkeitstheorie und Statistische
Physik
Das zentrale Problem der Statistischen Physik ist die Erklärung des makroskopischen Verhaltens der Materie aus dem mikroskopischen Verhalten ihrer Atome. Da solch eine Erklärung notwendigerweise stochastischer Natur ist, besteht eine fundamentale Beziehung zwischen Wahrscheinlichkeitstheorie und Statistischer Physik. Wahrscheinlichkeitstheoretische Methoden sind unverzichtbar für die Statistische Physik, und Konzepte der letzteren (wie z.B. die Entropie) sind zum festen Bestandteil der Stochastik geworden. Räumliche stochastische Modelle bilden das geeignete wahrscheinlichkeitstheoretische Werkzeug zur Untersuchung von interagierenden Vielteilchensystemen im thermodynamischen Gleichgewicht. Die zeitliche Entwicklung solcher Systeme kann durch stochastische Prozesse modelliert werden. Ein konkreter Gegenstand aus der Gleichgewichtstheorie, der im Graduiertenkolleg untersucht werden soll, ist die Theorie der großen Abweichungen, welche die Fluktuationen extensiver Größen beschreibt und auch für das klassische Problem der Äquivalenz der Gibbsschen Ensembles von Bedeutung ist. Dieser Gegenstand soll u.a. auch für ungeordnete Systeme wie die sogenannten Spingläser betrachtet werden. Über die Ergodentheorie, insbesondere die zufälligen dynamischen Systeme, und die zufälligen Schrödinger-Operatoren ergibt sich von hier eine Querverbindung zu Schwerpunkt A.
In der Nichtgleichgewichtstheorie soll die Herleitung von Transportphänomenen aus mikroskopischen Teilchenprozessen im hydrodynamischen Limes im Vordergrund stehen. Notwendig hierzu sind Methoden aus der Theorie der Diffusionsprozesse und der großen Abweichungen. Über die dabei resultierenden Transportgleichungen ergibt sich auch hier eine enge Beziehung zu Schwerpunkt A.
Außer den eben genannten Fragestellungen, die der klassischen Statistischen Physik angehören, sollen auch Grundlagenprobleme der Quantenphysik Gegenstand des Graduiertenkollegs sein. Hier konzentriert sich das Interesse einerseits auf die Untersuchung quantenstatistischer Transportphänomene, andererseits aber auch auf Grundlagenprobleme wie die Beschreibung des Meßprozesses in der Quantentheorie durch Spektralmaße von Operatoren. Über die algebraischen Strukturen von Operatoren ergibt sich ein natürlicher Bezug zu Schwerpunkt C. In diesen Problemkreis gehört auch die Frage nach der geeigneten Formulierung einer relativistischen Quantentheorie. Die damit verbundene Untersuchung von Raum-Zeit-Symmetrien und der allgemeinen Struktur der Raum-Zeit ist auch Gegenstand der Schwerpunkte B und C.
In der Beschreibung der Schwerpunkte in 4.3 werden Bezüge und Vernetzungen genannt, die sich in natürlicher Weise aus den Forschungsprogrammen ergeben. Diese Bezüge werden zum Teil im Graduiertenkolleg eingehend verfolgt und untersucht, so daß sich eine rege Zusammenarbeit zwischen den Schwerpunkten ergibt. Diese manifestiert sich z.B. in der Auswahl der auswärtigen Gäste, in der Teilnahme an Lehrveranstaltungen im Programm des Graduiertenkollegs sowie an Frühjahrs- und Herbstschulen (vgl. 8.1) und in der Betreuung und Beratung der Doktoranden. Dabei ist das Zusammenwirken aufgrund der speziellen Interessen der beteiligten Hochschullehrer zwischen den Schwerpunkten A und D einerseits und B und C andererseits enger als bei den anderen möglichen Kombinationen.
Die grundsätzlichen Ziele des Studienprogramms sind es,
Zur Erreichung dieser Ziele dienen folgende Veranstaltungen:
Auf Anregung von Herrn Prof. Dr. J. Wess wurde im Jahre 1999 ein Ergänzungsantrag gestellt mit dem Ziel, Doktoranden und Postdoktoranden aus dem ehemaligen Jugoslawien zu fördern. Der Antrag wurde zügig behandelt und genehmigt, so dass für die zwei abschließenden Jahre April 2000 - März 2002 (und zum Teil in der Auslauffinanzierungszeit April 2002 - März 2003) zusätzlich zu den vorhandenen Stellen zwei Postdoktoranden und ein Doktorand aus Serbien, Montenegro und Kroatien gefördert werden konnten.
Auslauffinanzierung)
BAKOVIC, Igor (Wess/Schottenloher)
DIDT, Daniel (Schneider)
ELSNER, Karen (Maison)
GROSSKINSKY, Stefan (Spohn)
HOBST, Daniela (Pareigis)
PICKL, Peter (Dürr)
SCHMIDT, Alexander (Wess)
zugeschnittenen Methoden angehen. Gegenwärtig schälen sich immer mehr Problemkreise heraus, die sich einem solchen direkten Zugang entziehen, und für deren Lösung eine größere theoretische Durchdringung des Umfelds erforderlich ist. Beispiele sind etwa
Auch von Seiten der Modelltheorie sind für die genannten Anwendungsfelder wichtige Impulse zu erwarten. Die Entwicklung eingebetteter Systeme stellt eine große Herausforderung für die Sicherstellung der Korrektheit dar, da in allen Anwendungen hohe Anforderungen an die Korrektheit gestellt werden. Die bisher verwendeten Entwurfstechniken (Statecharts - Statemate, SDL, Lotos, Esterelle usw.) haben allesamt gemein, dass sie semantisch nicht voll erschlossen und logisch nicht sauber fundiert sind. Eine vollständige mathematische und insbesondere logische Fundierung ist jedoch nach Stand der Technik heute möglich und, um Korrektheit in hochkritischen Anwendungen sicherzustellen, auch unumgänglich.
Die auftretenden theoretischen Fragestellungen decken das weite Spektrum der logischen Teildisziplinen (Beweistheorie, Modelltheorie, Rekursionstheorie, Komplexitätstheorie, Typentheorie etc.) ebenso ab wie die der Informatik (model checking, formale Spezifikation und Verifikation, Computer-Aided Reasoning u.v.m.). Die theoretische Tiefe der anstehenden Aufgaben und der benutzten Methoden macht das Forschungsgebiet hervorragend geeignet für Promotionsvorhaben wie sie in den Forschungsprogrammen der Dozenten und Kollegiaten beschrieben sind. Die Integration und Vernetzung der verschiedenen Aktivitäten in dem weitgefächerten Gebiet ist das besondere Anliegen unseres Graduiertenkollegs.
Zwischen diesen Arbeitsgebieten gibt es offensichtlich zahlreiche Überlappungen. Die sich daraus ergebenden Kooperationen sind erklärtes Ziel des Graduiertenkollegs und haben in der Vergangenheit einen besonderen Niederschlag in den folgenden Schwerpunktbereichen gefunden.
The Department "Institute for Informatics" is highly reputed for its research. In the first ranking of Computer Science departments at German Universities performed by Professor Dr. B. Nebel in 2001 (cf. Informatik Spektrum, August 2001), the Institute for Informatics scored rank 2 out of 50 institutions in the category average number of research publications per professor and scored rank 4 out of 50 in the category average number of citations in CITESEER per professor.
During the last few years, several prestigious research awards and research employment offers at well-known institutions were received by members of the Institute for Informatics: ACM-SIGMOD Best Paper Award 1997, IFIP/IEEE Dan Stokesberry Memorial Award 2001, offer to lead a research group at AT&T research labs, offer for a full professorship (chair) at the United Nations University.
Furthermore, 3 professors of the Institute for Informatics are members of the Informatics Commission of the Bavarian Academy of Sciences and 2 professors of the Institute for Informatics are co-ordinators of (thematics or excellence) networks of the EU commision. An additional indicator for the high quality research of the department is the high and yearly increasing amount of external research funds spent by the Institute for Informatics: 933.000 Euro in 2000, 1.106.000 Euro in 2001, and 1.521.000 Euro in 2002.
The Institute of informatrics consists of 7 units:
Prof. Dr. Heinz-Gerd Hegering, full professor
Prof. Dr. Claudia Linnhoff-Popien, full professor
| 1998 | 1999 | 2000 | 2001 | 2002 | 2003 | ||
| Full professor(s): | 1 | 1 | 1 | 1 | 1 | 1 | |
| Associate professor(s) : | 1 | 1 | 1 | 1 | 1 | 1 | |
| Research and teaching assistant(s) (*) | 5 | 5 | 5 | 5 | 6 | 6 | |
| Research assistant(s) (without teaching duties) (**) employed at (***) | LMU | 3 | 4 | 4 | 5 | 5 | 8 |
| LRZ/TUM | 4 | 4 | 4 | 4 | 4 | 4 | |
| elsewhere | 1 | 1 | 1 | 1 | 5 | 2 | |
(*) "Wissenschaftlicher Mitarbeiter auf Landesstelle mit
Lehrverpflichtung"
(**) "Wissenschaftlicher Mitarbeiter aud Drittmittelstelle oder
mit Promotions- oder Postdoc-Stipendium ohne
Lehrverpflichtung"
(***) LMU: University of Munich, LRZ: Leibniz Supercomputing
Centre, TUM: Technische Universität München
In addition to his profession as a professor, Dr. Hegering is chairman of the board of directors at the Leibniz Supercomputing Centre of the Bavarian Academy of Sciences.
The following topics characterise the main focus of research and teachings:
At present, especially the research areas below are addressed that also lead to several long-term third-party funded projects.
The research work of this unit is strongly international oriented. Prof. Hegering for example received the IFIP/IEEE Dan Stokesberry Memorial Award "for his outstanding contribution to the field of Integrated Network Management" in 2001.
As a consequence of global-wide networking and new forms of tele-cooperations the commercial relevance of networked systems, distributed applications and distributedly provided IT-services is increasing. As one of the first groups in Germany this research unit is engaged in questions of the network-, system-, application-, enterprise- and service-management. Research is done in tight cooperation with international developers and users and in consideration of standards and best-practice-guides like ITIL, eTOM and others. This topic is intertwined with economic questions of business process modeling and the foundation of virtual organisations.
Service management mainly is concerned with the development of new concepts and a universal architecture (service-management platform) to form the basis for an all-embracing service-management. Subquestions are the specification of service dependencies, the development of a universal service model, the modeling of customer-provider-interactions at the service access interface and the modeling of service level agreements (SLAs).
Policy-based Management. Policies enable the transition from purely imperative management approaches to a declarative paradigm and therefore provide better support for distributed management. These policy-based techniques are predestinated for highly dynamic and mobile environments. Research is done with cooperation partners to analyse how far these concepts are suitable for the management of, e.g., mobile networks of the 4th generation. Another research topic is the development of a methodology to support business and management processes using policy techniques with the objective to realize them automated in concrete management architectures. With the distributed specification of policies and their refinement into operational policies (machine-executable) conflicts can occur. There are two basic approaches to treat these conflicts: The preventive detection of potential conflicts and the reactive solution of conflicts that already have occurred. For both possibilities formal IT-management-models are analysed and specified to find invariances that are applicable to conflict-treatment.
Management of customer-provider-interactions. Utilising and provisioning of services as well as their quality are negotiated between customer and provider in a service level agreement (SLA). The unambiguous specification of quality attributes that are monitored and enforced during the operation of a service is a precondition for meaningful SLAs. Therefore formal and all-purpose description and modeling techniques based on existing standards are developed. In today's systems customers increasingly demand a technical interaction-interface to the management-system of the service-provider. At the so called customer-service-management (CSM) interface the customer shall be granted a uniform, but client-specific, view to the service, its current state, its current quality of service (QoS) as well as to the reporting of the service. Above all the customer shall be able to adapt the service to his current needs within the limits that were negotiated before.
In order to realise a CSM the provider has to change its component-based management system, thereby paving the way for a customer-based and service-specific paradigm. This leads to new questions. To identify the quality of a service customer-specific parameters have to be declared and monitored. Therefore, methods have to be found to map the quality-of-service-parameters on the several layers of the management hierarchy, which is to map customer-specific QoS-parameters on system- and component-specific QoS-parameters and vice versa. Today complex services are not offered by a single provider but by a chain of several providers, therefore mappings on the several QoS-architectures of each provider are also necessary (difficulty of mapping in the provider hierarchy).
Similar difficulties exist within the fault- and problem-management. The provider is forced to consider his infrastructure and his components from the point of view of the customer. In this context the question occurs how component-driven error-messages (events) could be associated to a specific service or customer. On the other hand there is the need to find the causative component as fast as possible after receiving an error report (trouble ticket) from a customer concerning a specific service. Both mappings, bottom-up as well as top-down are subject of research work.
Architectures and Concepts for the Support of process-oriented Service-Management. Process-oriented Service-Management allows a technology-independent view to the operational realisation of IT-management. Present tools and architectures for the support of IT-management either deal with the monitoring and controlling of the technical infrastructure exclusively or only with single processes of service-management. Up to now, there are no architectures or concepts that would be able to integrate these tools and to offer an all-embracing support of all processes. Though the TeleManagement-Forum within NGOSS (New Generation Operations Systems and Software) that includes eTOM is engaged in this problem field, there is no respective framework in sight for the ITIL that is more widely-spread in Europe. The definition of appropriate information models as well as the difficulties concerning workflow-controlling and communication among distributed management-applications are subject of the research work of this unit.
Publications on the above-mentioned and/or related issues are [3], [5], [9], [11], [13], [14], [15], [16], [17], [19], [21], [22], [23], [24], [25], [31], [32], [33], [35], [36], [38], [39], [41], [42], [43], [44], [45], [48], [49], [50], [52], [53], [54], [55], [57], [58], [59], [60], [62], [65], [67], [69], [70], [71], [72], [73], [74], [78], [81], [82], [83], [85], [87], [88], [89], [90], [91], [95], [99], [104], [106], [108], [109], [113] and [118].
| Contributing researchers: | Heinz-Gerd Hegering, Michael Brenner, Vitalian Danciu, Gabi Dreo-Rodosek, Christian Ensel (until June 2002), Markus Garschhammer, Boris Gruschke (until November 1999), Andreas Hanemann, Rainer Hauck (until July 2001), Stephen Heilbronner (until January 2000), Thomas Kaiser (until April 1999), Alexander Keller (until December 1998), Bernhard Kempter, Michael Langer (until March 2001), Stefan Loidl, Michael Nerb (until March 2001), Thomas Paintmayer (until May 1999), Igor Radisic (until February 2003), Helmut Reiser, Harald Roelle, Martin Sailer, Holger Schmidt (until July 2001), David Schmitz, Norbert Wienold (until June 2001) |
| Cooperations: | BMW Group, T-Systems, Siemens ICN, Siemens CT, Cisco, IBM Research Labs Rüschlikon, INRIA Nancy,Hewlett-Packard Research Labs Bristol, Frauenhofer Fokus, German Research Network (DFN), DZ Bank, RWTH Aachen, Imperial College London, University of Pretoria |
| Research Grant: | Seven full-time researcher positions by DFN, Siemens ICN and T-Systems |
An essential ancillary condition for a successful distributed management is the security of the management system. Management systems are increasingly operated inter-organisational. Because of the vital relevance of a management system for the operation of distributed IT-infrastructures concepts are necessary which incorporate security aspects in an all-embracing way. Threat analyses or security risk analyses deliver requirements a secure distributed management system has to meet. Research is done on how a system that is operated by several organisations could integrate all aspects of security (e.g. authentication, authorisation, access control, abuse treatment etc.) and what services are needed therefore. Subquestions are among other things how the security of critical management functions could be increased with the use of federative security concepts and extended authorisation mechanisms. And how single management tasks could be delegated in an efficient and secure way. All this aims towards a security architecture for federative and interorganisational management systems.
Publications on the above-mentioned and/or related issues are [8], [10], [28], [34], [37], [68], [73] and [92].
| Contributing researchers: | Heinz-Gerd Hegering, Helmut Reiser, Gerald Vogt (until March 2003) |
| Cooperations: | Dr. Matthias Scherf, Genomatics; Prof. Dr. Thomas Seidl, RWTH Aachen; Prof. Dr. Thomas Meitinger and Dr. Holger Prokisch, Institute of Human Genetics, GSF National Research Center, Neuherberg, Germany. |
| Cooperations: | IBM Research Labs Zürich, Bundesamt für Sicherheit in der Informationstechnik (BSI) |
| Research grant: | One full-time researcher position funded by the Graduiertenkolleg TU Munich |
IT-outsourcing is the relocation of IT-services to external service providers who are not part of the own company. The provisioning of complex application services via a network (ASP) holds particular challenges for the IT-management. In addition to classic difficulties of QoS-management, these challenges also include the addressing of security concerns on the part of the customer and the provisioning of controlling tools for the customer (e.g. for ad-hoc ordering of standard services or requesting accounting-information), because outsourcing leads to the customer losing direct control over service provisioning. Additional questions concerning this topic arise against the background of increasing trends towards a partial dynamic composition of services (e.g. with Web Services) across several domain boundaries.
Publications on the above-mentioned and/or related issues are [29], [30], [36], [40], [56], [62], [86], [93], [94], [103], [105], [111], [115], [119], [120] and [121].
| Contributing Researchers: | Heinz-Gerd Hegering, Markus Garschhammer, Michael Langer (until March 2001), Michael Nerb (until March 2001), Harald Roelle |
| Cooperations: | German Research Network (DFN), BMW Group, T-Systems |
| Research Grant: | Two full-time researcher positions funded by the German Research Network (DFN) |
New challenges arise because of mobility and context-awareness. Users access services of multiple providers with manifold end devices via heterogeneous access networks (e.g. GSM/UMTS, WLAN, Bluetooth etc). This leads to highly dynamic relations between involved service components and organisations. To achieve an efficient service management in mobile environments, decentralised and self organising management approaches are increasingly considered. Together with industry partners this unit currently develops a policy-based management system tailored to the needs of mobile networks. Due to the heterogeneity of end devices and access networks, services as well as content must be adaptable. Also under examination is how adaptation functionality could be integrated in so called Content Delivery Networks and be combined with the distribution of content and services.
As a consequence of mobility new applications become possible. One example are context-aware services that are adapted depending on context-information like the user's position or current profile information. In consideration of aspects like efficiency, security, accounting and fault tolerance, strategies and infrastructure services are developed that allow the provisioning, processing and using of context information in cellular and ad-hoc-networks. Because context-aware services will apparently affect the economy and society, this unit is engaged in an interdisciplinary project with the business administration and communication science departments. Here, in addition to technical issues, business models and their influence on existing value chains in the media industry are examined as well as the users' acceptance and the effects of context-aware services on society.
Publications on the above-mentioned and/or related issues are [1], [2], [6], [7], [47], [51], [77], [93], [107], [110], [114] and [122].
| Contributing Researchers: | Heinz-Gerd Hegering, Claudia Linnhoff-Popien, Thomas Buchholz, Vitalian Danciu, Caroline Funk, Iris Hochstatter, Michael Krause, Axel Küpper, Michael Samulowitz (until July 2002), Michael Schiffers, Thomas Strang |
| Cooperations: | Prof. Alexander Schill, TU Dresden; Björn Schiemann, Siemens Corporate Technology; Andre Varney, Apollis Interactive; Prof. Arnold Picot, LMU; Prof. Hess, LMU, Prof. Brosius, LMU; Prof. Wirth, ETH Zürich |
| Research Grant: | Two full-time researcher positions funded by the BMBF within the project "Internetökonomie". Two full-time researcher positions funded by the Bayerische Staatskanzlei within the project "Kontextsensitive Dienste für globale UMTS-Netze" |
New Internet and web-based technologies lead to new chances and difficulties in teaching. This unit examines concepts of virtual learning and teaching for lectures as well as for practical courses. The lecture "Telecommunication Systems" was broadcasted between Munich and the RWTH Aachen (Aix-la-Chapelle) and the lecture "Computer Networks" between Munich and Garching via interactive video-conferences.
An interactive, multimedia-based practical course was offered for the first time in the summer term 2003 in cooperation with the RWTH Aachen. Through distributed access to learning material via the Internet the students gained more flexibility with respect to the time and place of their learning activities. This flexibility requires an integrated organisation of students, teachers and learning material. Current research work concentrates on the support of practice-related learning with simulation-based learning material, which emulates the highly dynamic behaviour of communication networks. This learning material is offered via a web-based interface for a distributed, interactive and self organising learning.
Publications on the above-mentioned and/or related issues are [4], [12], [46], [64], [76] and [102].
| Contributing Researchers: | Claudia Linnhoff-Popien, Mang Li |
| Cooperations: | Prof. Dr. Otto Spaniol, RWTH Aachen; Prof. Dr. Uwe A. Michelsen, RWTH Aachen; Michael R. Kunzi, Cisco Systems |
| Research Grant: | One full-time researcher funded by the BMBF within the project "Entwicklung und Evaluierung eines Virtuellen Informatik Praktikums" |
Since 1998, the following theses have been completed by the members of the Unit:
Prof. Dr. Hans-Peter Kriegel, full professor
Prof. Dr. Christian Böhm, associate professor
| 1998 | 1999 | 2000 | 2001 | 2002 | 2003 | |
| Full professor(s): | 1 | 1 | 1 | 1 | 1 | 1 |
| Associate professor: | 1 | 1 | 1 | |||
| Research and teaching assistant(s) (*) | 7 | 7 | 7 | 8 | 8 | 8 |
| Research assistant(s) (without teaching duties) (**) | 1 | 2 | 2 | 3 | 3 | 3 |
(*) "Wissenschaftlicher Mitarbeiter auf Landesstelle mit
Lehrverpflichtung"
(**) "Wissenschaftlicher Mitarbeiter aud Drittmittelstelle oder
mit Promotions- oder Postdoc-Stipendium ohne Lehrverpflichtung"
Hans- Peter Kriegel and Christian Böhm received the internationally prestigious "SIGMOD Best Paper Award 1997" for the publication and prototype implementation "Fast Parallel Similarity Search in Multimedia Databases" together with three members of their research team. This prize ranks as one of the highest awards for research in database systems.
The projects described in more detail below led to several grants by the Deutsche Forschungsgemeinschaft (DFG), by the Bundesministerium für Bildung und Forschung (BMBF) and by industrial partners.
Modern methods in application domains such as molecular biology, astronomy, geography, etc. produce a vast amount of data. Since all this data can no longer be managed without the help of automated analysis tools, there is an ever increasing need for efficient and effective data mining methods to make use of the information contained implicitly in that data. One of the main tasks of data mining is clustering which is intended to help a user discovering and understanding the natural structure or grouping in a data set. Clustering real-world data sets is often hampered by the so-called curse of dimensionality, since many real-world data sets consist of a very high dimensional feature space. In this case, traditional clustering algorithms often fail to generate meaningful results. We are currently working on subspaces clustering algorithms and advanced feature selection methods for clustering high-dimensional data.
Web content mining is the application of data mining techniques to automatically extract relevant content from the world wide web. While many established methods are targeted to retrieve relevant web pages, there are various applications which target whole web sites. A web site is a set of web pages serving a common purpose, e.g. to represent an organisation. We develop techniques for classification, focused crawling and clustering to enable users to query the web for specific web sites.
Another important aspect is the development of data mining algorithms for complex objects. Since representing complex objects as single feature vectors might loose vital information, we develop methods that are capable to handle sets and tuples of feature vectors. Applications of this approach are biomolecular databases and CAD-databases which provide multiple representations for the data objects.
Traditional KDD methods require full access to the data which is going to be analysed. All data has to be located on a single site. Nowadays, large amounts of heterogeneous, complex data reside on different, independently working computers which are connected to each other via local or wide area networks, e.g. mobile and sensor networks, supermarket chains, and cannot be made available in a centralised location due to bandwidth limitations, scalability problems or security/privacy considerations. We develop methods for analysing data in a distributed fashion. Thus, the data is analysed locally, where it has been generated and stored. Afterwards, we combine these local models to a global one. For large data sets, scaling up the speed of the KDD process is crucial. Thus we further develop techniques for parallel knowledge discovery.
Publications on the above-mentioned and/or related issues are [1], [5], [6], [8], [13], [15], [16], [18], [19], [20], [21], [22], [23], [27], [28], [29], [32], [33], [34], [36], [37], [43], [45], [48], [50], [51], [52], [53], [56], [57], [61], [63], [65], [66], [68], [70], [71], [72], [73], [75], [76], [77], [78], [79], [87], [88], [89], [90], [91], [96], [97], [98], [99].
| Contributing researchers: | Hans-Peter Kriegel, Christian Böhm, Mihael Ankerst (until March 2001), Stefan Berchtold (until July 1997), Bernhard Braunmüller (February 2001), Markus Breunig (until February 2001), Martin Ester (until October 2001), Eshref Januzaj, Karin Kailing, Daniel Keim (until March 1998), Florian Krebs (until August 2001), Peer Kröger, Martin Pfeifle, Jörg Sander (until August 2001), Stefan Schönauer, Thomas Seidl (until August 2002), and Matthias Schubert, Xiaowei Xu (until June 1998) |
| Cooperations: | Prof. Dr. Jiawei Han, University of Illinois, Urbana Champaign, USA; Prof. Dr. Raymond Ng, University of British Columbia, Vancouver, Kanada; Dr. Mihael Ankerst, Boeing, Seattle, USA; Prof. Dr. Martin Ester, Simon Fraser University, Vanouver, Kanada; Prof. Dr. Jörg Sander, University of Alberta, Edmonton, Kanada. |
| Research grant: | one full-time researcher position funded by the Deutsche Forschungsgemeinschaft (DFG) within the project "Methods for cluster structure analysis and KDD database primitives for efficient (semi-) automatic knowledge extraction in very large high-dimensional databases" |
The explosive growth of digital data has opened great challenges to the area of E-business. A family of technologies, summarized as "Information Filtering", emerges to assist users to find desired information from large amount of available data. Applications include recommender systems, image retrieval, spam email filtering, news filtering, personalized web browsing and others. It has become a vivid research area that focuses on learning users' profiles of information need. Existing approaches either infer individuals' profiles by exploring their connections to other users, known as collaborative filtering, or learn the profile from the descriptive contents of examples given by the users, known as content-based filtering. This project studies various novel statistical learning approaches to modeling user profiles, namely collaborative filtering, content-based filtering and hybrid filtering.
Publications on the above-mentioned and/or related issues are [51], [52], [69], [76], [85], [90], [91].
| Contributing researchers: | Hans-Peter Kriegel, Martin Ester (until October 2001), Xiaowei Xu (until June 1998), Zhao Xu, Kai Yu, Shipeng Yu |
| Cooperations: | Prof. Dr. Bernd Schürmann, Dr. Volker Tresp, Siemens AG München Corporate Technology |
| Research grant: | three doctoral scholarships funded by Siemens AG, Corporate Technology |
Modern database applications are characterised by two major aspects: the use of complex data types with internal structure and the need for new data analysis methods. The focus of database users has shifted from simple queries to complex analysis of the data, known as knowledge discovery in databases. Important tasks in this area are the grouping of data objects (clustering), the classification of new data objects or the detection of exceptional data objects (outlier detection). Most algorithms for solving those problems are based on similarity search in databases.
On the other hand, similarity search is also directly important in many application domains. In molecular biology and pharmaceutics, finding geometrically similar or geometrically complementary molecules is essential for computer-aided drug design. In engineering, enormous cost reductions are possible by the identification of similar parts and a following reduction of part diversity.
All this makes efficient similarity search in large databases of structured objects an important basic operation for modern database applications. Consequently, similarity search plays an important role in several of our research projects and is investigated from different view points in those projects.
In the past few years, we developed several successful similarity search methods for different applications. A very general approach is based on a histogram representation of objects and the use of quadratic form distance functions. We applied this approach in such diverse application domains as image retrieval, bio informatics and CAD databases. The use of the quadratic form distance function allows the flexibility to adapt the similarity search method to the needs of different applications and users. This extends even to a point where the user is able to adapt the similarity measure between two queries without loosing interactive response times. The necessary performance is achieved through several generally applicable techniques like multi-step query processing architectures, an optimal k-nearest-neighbour algorithm, dimension reduction and efficient index structures.
We also developed efficient similarity search methods for graph- and tree-structured data. The edge matching distance for such data uses a matching between edge sets to derive a measure for the similarity of graph structures. It is characterised by a moderate computational complexity and high adaptability to user and application needs. Additionally, a multi-step query processing architecture for this measure was created to ensure good query response times. For tree-structured data, a similar query processing architecture was developed, also showing very good performance results.
All our similarity search techniques have been implemented prototypically and tested with data from various real-world applications. Additionally, several demos of the techniques are available on our web site.
Publications on the above-mentioned and/or related issues are [2], [3], [7], [9], [11], [12], [25], [30], [33], [36], [39], [40], [42], [47], [54], [55], [57], [63], [64], [65], [83], [86], [88], [93], [94], [103], [104], [105].
| Contributing researchers: | Hans-Peter Kriegel, Christian Böhm, Mihael Ankerst (until March 2001), Bernhard Braunmüller (until February 2001), Stefan Brecheisen, Markus Breunig (until February 2001), Eshref Januzaj, Karin Kailing, Daniel Keim (until March 1998), Florian Krebs (until August 2001), Peer Kröger, Martin Pfeifle, Marco Pötke (until September 2001), Jörg Sander (until August 2001), Stefan Schönauer, Matthias Schubert, Thomas Seidl (until August 2002) |
| Cooperations: | Dr. Mihael Ankerst, Boeing, Seattle, USA; Prof. Dr. Thomas Seidl, RWTH Aachen. |
The most well-known method for similarity search in various application domains such as CAD databases, time sequence analysis, molecular biology etc. is the so-called feature transformation. This transformation associates the objects of the database with vectors of numerical features which characterise the objects. Similarity search is thus translated into a search for neighbouring vectors in a usually high-dimensional feature space. The specific advantage is that using a single methodology, i.e. neighbourhood search upon feature vectors, a broad variety of application domains can be handled. One of the most important index structures for neighbourhood search in low dimensional spaces, the R*-tree, has been developed in the LMU database group. Our extension for high-dimensional spaces, the X-tree has become one of the milestones and yardsticks for similarity search. Due to a number of effects commonly referred to as the "curse of dimensionality", even the performance of the X-tree and other specialised index structures for high dimensions deteriorates when facing feature spaces of too high dimensionality. To gain a deep understanding of the related problems, a cost model for query processing based on index structures like the R*-tree or the X-tree was developed in a first step. This cost model was then used to identify numerous parameters for optimisation. A number of optimisation techniques such as the dynamic block size optimisation, optimal dimension assignment, optimal index scan techniques, and an optimising data compression technique for index structures have been proposed. The IQ tree, into which we integrated many of these techniques, has become one of the most efficient index structures today.
Publications on the above-mentioned and/or related issues are [4], [10], [12], [14], [22], [23], [24], [31], [33], [35], [36], [39], [40], [41], [42], [45], [46], [47], [49], [50], [56], [57], [63], [64], [65], [66], [68], [70], [71], [72], [75], [77], [78], [80], [81], [82], [84], [95], [101], [102].
| Contributing researchers: | Hans-Peter Kriegel, Christian Böhm, Bernhard Braunmüller (until February 2001), Daniel Keim (until March 1998), Florian Krebs (until August 2001), Peter Kunath, Martin Pfeifle, Marco Pötke (until September 2001), Matthias Renz, Matthias Schubert, Thomas Seidl (until August 2002) |
The development, design, manufacturing and maintenance of modern engineering products is a very expensive and complex process. Today, thousands to millions of CAD files of a car or an air plane may occupy terabytes of distributed secondary and tertiary storage. The main objective of this project is to find techniques to manage effectively and efficiently huge enterprise-wide amounts of spatial data and to evaluate them in an industrial environment. As a main application, we focus on collision detection and within-distance queries for digital mockup and on similarity search in large CAD databases. For an efficient query processing, conservative approximations of the CAD models are stored in a database in order to support geometric queries like collision or within-distance queries in a multi-tiered fashion. A selective and complete filtering mechanism determines only a small amount of the CAD parts to be downloaded from the database into main memory for exact geometric analysis. The efficiency is in particular realized by index structures we developed, which can be easily integrated into industrial strength object-relational database systems. To reduce the query times when dealing with high resolution spatial data, we apply data compression techniques to lower the high I/O cost. Special care is taken to find a good compromise between I/O- and CPU cost. In addition, we analyse which degree of redundancy is most suitable for high resolution spatial objects. Collision detection and within-distance queries, which are very important for digital mockup (DMU), can be carried out interactively using our approach, instead of running time-consuming batch jobs on the whole database overnight. Furthermore, our approach for the efficient management of spatially extended objects can be applied on many other modern database applications including geographic information systems (GIS), medical imaging, computer- aided design (CAD), molecular biology or time-critical virtual reality applications, e.g. haptic rendering. Our industrial partners Boeing, Volkswagen and the German Aerospace Centre (DLR) use and evaluate the methods developed by us.
Publications on the above-mentioned and/or related issues are [3], [4], [9], [24], [38], [58], [59], [60], [62], [67], [74], [82], [83], [84], [86], [92], [94], [95], [100], [101], [102], [103].
| Contributing researchers: | Hans-Peter Kriegel, Christian Böhm, Mihael Ankerst (until March 2001), Stefan Brecheisen, Peer Kröger, Peter Kunath, Martin Pfeifle, Marco Pötke (until September 2001), Matthias Renz, Matthias Schubert, Thomas Seidl (until August 2002) |
| Cooperations: | Dr. Andreas Müller, Volkswagen AG; Dr. Bob Abarbanel, Boeing; Prof. Dr. Gerd Hirzinger and Carsten Preusche, DLR; Dr. Marco Pötke, sd&m. |
Modern analysis methods in molecular biology produce a tremendous amount of data. To make the full use out of this data, automatic tools for knowledge discovery and data analysis are needed. Our project "Molecular Bioinformatics" tries to address this task of supporting the automatic analysis of molecular biology data. The first sub-project aims at supporting the 1:n protein docking problem, i.e. given a query protein P find all proteins in a database that interact physically with P. A typical application is the prediction of the effect of specific drugs and their side effects to the human immune system. To predict the docking partners, potential docking sites are located on the surface of all proteins in a database for a given query molecule. Specific shape descriptions of these sites enable an efficient search in the database. The second sub-project deals with the cluster analysis of gene expression data. Micro array chip technology allows monitoring the expression level of thousands of genes synchronously. From the expression level of a gene, biologists are able to make implications about the amount of the gene product, it codes for (e.g. a protein). To analyse which genes are co-expressed under which conditions, tools for clustering are used. Since the raw data is very high-dimensional and genes typically cluster (are co-expressed) differently under varying conditions, the development of new clustering methods is required, for example algorithms for subspace clustering.
Publications on the above-mentioned and/or related issues are [2], [9], [17], [25], [26], [30], [87].
| Contributing researchers: | Hans-Peter Kriegel, Mihael Ankerst (until March 2001), Karin Kailing, Peer Kröger, Stefan Schönauer, Thomas Seidl (until August 2002) |
| Cooperations: | Dr. Matthias Scherf, Genomatics; Prof. Dr. Thomas Seidl, RWTH Aachen; Prof. Dr. Thomas Meitinger and Dr. Holger Prokisch, Institute of Human Genetics, GSF National Research Center, Neuherberg, Germany. |
| Research grant: | two full-time researcher positions funded by the BMBF within the joint project: "Bioinformatics for the Functional Analysis of Mammalian genomes (BFAM)" |
During the report period, Hans-Peter Kriegel received an offer for the position of technology leader for the database group, AT&T Labs Research, USA, as well as offers for chairs in database systems with the Universität Augsburg and Cardiff University, GB. All offers were declined.
Prof. Dr. Martin Wirsing, full professor
Prof. Dr. Fred Kröger, associate professor
| 1998 | 1999 | 2000 | 2001 | 2002 | 2003 | |
| Full professor(s): | 1 | 1 | 1 | 1 | 1 | 1 |
| Associate professor(s) : | 1 | 1 | 1 | 1 | 1 | 1 |
| Research and teaching assistant(s) (*) | 6 | 6 | 6 | 7 | 7 | 7.5 |
| Research assistant(s) (without teaching duties) (**) | 10 | 9 | 14 | 13 | 11 | 7.5 |
(*) "Wissenschaftlicher Mitarbeiter auf Landesstelle mit
Lehrverpflichtung"
(**) "Wissenschaftlicher Mitarbeiter aud Drittmittelstelle oder
mit Promotions- oder Postdoc-Stipendium ohne Lehrverpflichtung"
See below for a more detailed table with researcher's names.
The research and teaching unit on Programming and Software Engineering (PST) is focussing on practical applications and theoretical foundations of object-oriented software engineering methods, new generation CASE systems, and new programming languages. A prime topic of the R&D activities of PST is the development of methods to improve software quality. Formal methods involve making mathematical models of software and deriving properties of the software from the models. This parallels practice in other areas of engineering, but is more difficult for software, due to the extreme complexity of modern software systems, and the unfamiliarity of engineers with the complex discrete mathematics required. The key for this task is to integrate semiformal engineering methods and notations, such as UML, with formal methods in order to provide a basis for mathematical analysis, validation, maintenance and verification of requirements, design specifications and software systems. A second central theme is the development of new techniques for object-oriented, constraint based, functional programming. This includes the study of foundational aspects of programming languages, such as semantics, correctness, specification construction, and design and development calculi as well as the application of these methods and principles to the design and implementation of software systems.
Actual application areas are concurrent and distributed programming, global computing, Internet programming and multimedia systems. PST has particular competence in object-oriented software development of mobile systems, in Java programming and the design of web based software systems. PST researchers have developed the first formal semantics for multi-threading in Java and the first provably correct code generation algorithm, generating concurrent Java programs from UML designs. Tools developed by PST comprise an adaptive hypermedia training system (SMEX-Web), a workflow driven CASE tool for software engineering (PUSH) and for Web applications (OpenUWE), a model checker for UML real-time state and interaction diagrams (HUGO), a semantic meta editor for diagrammatic languages (RECOPLA), the language EPKML for high level design and rapid development of electronic product catalogues, and an editor for visual construction of Multi-Agent-Systems (Agent Component).
Currently, PST is coordinating the FET Global Computing project AGILE on developing software architecture of mobile systems; in addition, PST runs several sponsored research projects on formal foundations of object-oriented software engineering (InOpSys), web and hypermedia engineering (MMiSS), software development for distributed and mobile systems, test generation from design, and the development of a distributed web-based decision support for global change of climate (GLOWA).
In the following the actual research topics are presented in more detail.
Publications on the above-mentioned and/or related issues are [15], [62], [49], [79], [47], [107], [10], [9], [25], [67], [23], [66], and [95].
Model checking is an established technique for fully automatic proving temporal properties of reactive systems. The main limitation of model checking applications is its restriction to finite state systems. By using abstraction techniques, this limitation can be ameliorated to a certain degree. Consequently, model checking techniques have also been applied more recently to checking software and programs.
In our research, we aim at applying model checking to software designs rather than implementations. On the one hand, software models in the early design phases tend to be more abstract and thus more amenable to model checking. On the other hand, development costs can be greatly reduced by finding errors as early as possible. We use the "Unified Modeling Language" (UML) as the software modeling notation. A tool, called HUGO, has been developed for translating UML state machines and UML interactions over UML static structures into off-the-shelf model checkers. For un-timed systems we use the model checker SPIN, for models involving timing annotations we employ UPPAAL. HUGO can be used for consistency checks between the dynamic views of UML interactions and UML state machines. Moreover, UML interactions can be viewed as defining desired or forbidden runs of collaborating UML state machines, thus defining properties in an easy-to-use language.
Development of the application-oriented tool HUGO is complemented by theoretical research in abstraction techniques. Predicate diagrams pave the way for specifying and proving abstractions of infinite state systems diagrammatically and thus directly lend themselves for inclusion in the UML-based model checking tool HUGO.
Publications on the above-mentioned and/or related issues are [27], [45], [46], [63], [84], and [83].
Work in the area of Temporal Logic of Actions (TLA) investigates a diagram-based formalism to verify temporal properties of reactive systems. Diagrams integrate deductive and algorithmic verification techniques for the verification of finite and infinite-state systems, thus combining the expressive power and flexibility of deduction with the automation provided by algorithmic methods. We show that the concept of predicate diagrams can be used to verify not only discrete systems, but also some more complex classes of reactive systems such as real-time systems and parametrised systems. We prove the completeness of predicate diagrams and study an approach for the generation of predicate diagrams. We develop prototype tools that can be used for supporting the generation of diagrams semi-automatically.
We have proposed an extension of Lamport's Temporal Logic of Actions by spatial modalities for the specification and verification of systems involving mobile code. Similar as in the Ambient Calculus due to Cardelli and Gordon, we model the topological structure of a mobile system as a tree whose edges are labelled by names. The mobility of objects is reflected in the dynamical reconfiguration of this tree. The main novelty of our approach is a study of notions of refinement appropriate for mobile systems that guide the selection of the spatial modalities of the logic. Specifically, we use names to refer to locations at arbitrary nesting depth rather than just below the current node. In this way, a single agent may be decomposed into a hierarchy of sub objects during a refinement step. A first paper on the logic and its use for the specification and refinement of mobile systems has been published at FASE 2003. The logic has meanwhile been applied to the specification of an active traceroute protocol, and we have proved the decidability of its satisfiability and model checking problems.
Publications on the above-mentioned and/or related issues are [27], [96], and [117]
Work in the area of theoretical foundations of systems specifications centres around the mathematical theory of formal models for systems specification, development, transformation and verification. We investigate the algebraic and coalgebraic foundations of specifications and their relationships. An interesting application is the use of domain theory for computing provably correct solutions of ordinary differential equations.
In the area of algebraic specifications and term rewriting we have designed Timed Rewriting Logic and the language Timed Maude for specifying real-time object-oriented systems and we have developed techniques for extracting programs from proofs of structured specifications.
Timed Rewriting Logic (TRL) is a new variant of Rewriting Logic, an algebraic formalism which allows the dynamic behaviour of systems to be axiomatised using rewrite rules. In TRL rewrite rules can be labelled with time constraints and this provides a means of reasoning about time elapse in real-time systems. The object-oriented specification language Timed Maude is based on TRL. Timed Maude has been successfully applied to the specification of several pragmatic modeling languages including Message Sequence Charts and SDL.
Extraction of Structured Programs Our technique for obtaining correct programs from proofs produces correct high-level programs from parameterised specifications written in a subset of CASL. The technique is based on a combination of standard constructive mathematical proof methods (due to Curry and Howard) together with structural modifications deriving from the proof calculi for structured specifications. Although programs extracted by the Curry-Howard process can be very cumbersome, we use a number of simplifications that ensure that the programs extracted are in a language close to SML.
The notion of coalgebras for a functor on a concrete category provides a general framework which captures various types of state transition systems and process calculi. On this level of generality, we are interested in formal logics, which can be used to specify and reason about coalgebraically modeled systems. Our line of research follows three main threads, detailed below:
Duality Theory This branch exploits the fact that coalgebras are the formal duals of algebras, and tries to build bridges between both areas. The problem lies in the fact that coalgebras over a category C are dual to algebras over the category C^op, which is in general different from C. Therefore, the goals are the formulation of coalgebras and their logics in a category-independent manner, and use this description to dualise results which are already known in the algebraic setting, e.g. Birkhoff's variety theorem.
Formal Systems for Coalgebraic Logics Here, we are investigating proof systems which can be used along with coalgebraic logic. This requires that the logic is equipped with a concrete syntax, and that the relevant proof rules are isolated. The focus of research here is soundness, completeness, and the expressiveness of the logics.
Specification Logics The portion of the research concerned with specification logics tries to apply the general results obtained in the other two branches to the problem of specifying properties of concretely presented state based systems. The main contribution lies in a synthesis of the algebraic / coalgebraic approach, which enables to reason about observable properties, while at the same time retains a certain amount of compositionality, described by the algebraic aspect of the system.
Apart from providing denotational semantics for programming languages, domain theory can also be used as a foundation for exact computation in analysis. The translation of classical theorems of analysis into domain theory then provides exact and provably correct algorithms for computations in analysis. This general paradigm is substantiated with case studies in several areas, including the provably correct solution of ordinary differential equations and the implicit function theorem.
Publications on the above-mentioned and/or related issues are [13], [1], [92], [105], [18], [14], [44], [73], [118], [106], [7], [36], [93], [119], [19], [120], [58], [59], [5], [6], [24], [41], [42], [68], [43], and [65].
The programming language Java combines the object-oriented programming paradigm with the concurrent programming paradigm through tightly integrated support for multi-threaded programming. In particular, Java features a full-fledged memory consistency protocol for synchronising the local working memories of threads and the main memory. Java thus poses the challenge to integrate the low-level description of the memory model with the high-level language constructs into a modular, operational semantics which is parameterised in the details of the memory protocol.
We defined a mathematical structure, called event spaces, that declaratively captures the memory model specification. An operational semantics for the Java language constructs is combined with the event spaces by memory action primitives. In fact, the parameterisability of the approach has been proved by showing that an optimisation by so-called prescient store action is correct with respect to the original memory model. Moreover, event spaces have been demonstrated to provide a rich mathematical structure, viz. forming a prime-algebraic, consistently complete Scott domain, that clearly expresses the concurrent features of Java.
The event-based, operational semantics of Java does not answer the question for specifying and proving properties of Java programs directly and sufficiently. On the side of event spaces the logic of configuration theories, which is currently under thorough investigation, can be used to reason on the properties of memory models. For object-orientation itself, object specifications define rules for recursive object behaviour generalising the approach by Abadi and Leino.
Publications on the above-mentioned and/or related issues are [3], [8], [16], [20], [38], and [33].
| Contributing researchers: | Hubert Baumeister, Pietro Cenciarelli (until 1999), Rolf Hennicker, Stephan Janisch, Alexander Knapp, Piotr Kosiuczenko (until 2003), Alexander Kurz (until 2000), Stephan Merz (until 2002), Dirk Pattinson, Axel Rauschmayr, Bernhard Reus (until 2000), Martin Wirsing, Julia Zappe (until 2003). |
| Co-operations: | Japan Institute of Science and Technology, JAIST (Kanazawa, Japan), Protos Software GmbH (München), Software Research Associates (Tokyo), Michel Bidoit (Ecole Normale Supérieure de Cachan), Manfred Broy (TU München), Pietro Cenciarelli (Università di Roma "La Sapienza"; since 1999), Corina Cirstea (Southampton, UK), John N. Crossley (Monash University Melbourne), Abbas Edalat (Imperial College, London), Clemens Kupke (Amsterdam, Netherlands), Stephan Merz (INRIA Nancy; since 2003), Jose Meseguer (Univarsity of Illinois at Urbana Champaign), Bernhard Möller (Universität Augsburg), Ataru Nakagawa (Software Research Associates, Tokyo), Tobias Nipkow (TU München), Iman Poernomo (Monash University Melbourne), Kokichi Futatsugi (Japan Institute of Science and Technology, JAIST) Wolfgang Reif (Universität Augsburg), Bernhard Reus (University of Sussex; since 2000), Toshiki Sakabe (Nagoya University, Japan), Mingsheng Ying (Tsinghua University, Beijing). |
Architecture-based approaches have been promoted as a means of controlling the complexity of system construction and evolution, namely for providing systems with the agility required to operate in turbulent environments and adapt very quickly to changes in the enterprise world. Recent technological advances in communication and distribution have made mobility an additional factor of complexity, one for which current architectural concepts and techniques are not prepared for.
As part of the EU project AGILE we have developed extensions to the UML to model mobile systems. In a first step an extension of UML activity diagrams was defined to model the location of objects and how mobility primitives change the location of objects. How the thread of control is affected by objects changing their location is visualised by using the extension of UML sequence diagrams. To model the dynamic evolution of the interaction structure between multiple objects, UML state charts were extended by multiple input queues and by the notion of localities.
A new logic, extending the Temporal Logic of Actions (TLA) by Lamport, has been proposed for considering spatio-temporal issues in the specification and the analysis of the behaviour of mobile systems. The logic extends TLA by providing operators for the description of properties of trees, as well as changes in their structure, intuitively representing the hierarchical structure of administrative domains.
In the research area of Agent-Technology we have introduced an new approach (AgentComponent) that integrates software component features into agent technology. Component concepts like encapsulation, interfaces, parametrisation have been found useful for integration. Using this technology combination agents grow up to be reusable and parametrisable in different contexts. According to this approach a visual editor ( AgentComponent Tool) for AgentComponents has been developed and is still in development. This editor is based on the JADE Java Agent Development Framework and enables the user to visually create Multi-AgentComponent-Systems. Moreover AgentComponent-Systems can be visually parametrised and changed during runtime. This is a very useful feature to test the designed AgentComponent-Systems without having to stop the whole application.
Also in the research area of Agent-Technology we have designed and implemented BABSy, an accounting system that helps automate payment in an agent-based electronic commerce environment.
Publications on the above-mentioned and/or related issues are [103], [104], [97], [117], [89], [39], [113], [116], [115], [126], and [124].
Web-Engineering is a new upcoming discipline that focuses on systematic, disciplined and quantifiable approaches to the cost-effective development and evolution of high-quality Web-based systems and applications. The discipline of Web Engineering tries to prevent a new software crisis that may occur due to the omnipresence of Web applications and their strong cross-dependency. Web Engineering incorporates some of the well-known and successful traditional software engineering principles and practices, adapting them to the particular requirements of Web applications. But Web Engineering is inherently different from software engineering in that it must also take many aspects of other disciplines into account, such as:
In our research, we aim at supporting the modeling of Web applications and the model-driven implementation of Web systems. We developed a methodology - called UWE (UML-based Web Engineering) - to cover the life-cycle of Web application development proposing an object-oriented and iterative approach based on the standard UML and the Unified Software Development Process. UWE uses “pure” UML notation and UML diagram types as far as possible for the analysis and design of Web applications, i.e. without extensions of any type. Special aspects of the Web domain are addressed defining a so called “lightweight” UML profile. The UWE profile includes stereotypes and tagged values defined for the modeling elements needed to model the different aspects of Web applications, such as navigation, presentation, Web processes and adaptation aspects. The UWE design approach for workflow-based Web applications consists of introducing specific process classes that are part of a separate process model with a clear interface to the navigation model.
To support the systematic design the CASE-tool ArgoUWE has been implemented for Web information systems and is currently being extended for workflow-based Web applications. ArgoUWE is integrated in the OpenUWE tool suite environment to achieve a model-driven generation of such Web applications.
Publications on the above-mentioned and/or related issues are [12], [31], [32], [56], [52], [53], [55], [85], [87], [90], [75], [86], [88], [110], [111], [109], and [123].
SmexWeb (Student Modeled Exercising on the World Wide Web) is a modular framework for the development of adaptive Web-based systems. An instance of Smex Web has been implemented to give students the possibility to practice and deepen their knowledge in a time- and local independent fashion. At present SmexWeb offers a lesson on the topic of EBNF (Extended Backus-Naur Formalism).
The aim of the MMiSS project is to set up a multimedia Internet-based adaptive educational system, covering the whole subject of Safe Systems. The area of "Safe Systems" has in the last few years become increasingly important. Software is increasingly used to control security-critical embedded systems, in aeroplanes, spaceships, and cars; and electronic trading over the Internet, with its associated security risks, is rapidly expanding. All this requires qualitatively and quantitatively better training in Safe Systems. To go with the planned deployment at universities, a number of well-known German companies have already expressed, through the various industrial contacts of the project partners, an interest in measures for further training in their own workforce.
Publications on the above-mentioned and/or related issues are [4], [11], [26], and [112]
An important issue of global climate change concerns changes in the hydrological cycle and the investigation of ways for sustainable future water use. To improve the quality of predictions and global change simulations one has to take into account that many processes of various fields are involved and influence each other. Hence, a key challenge for integrative simulation and analysis is the integration of the various simulation models of the different disciplines ranging from natural sciences, such as meteorology, hydrology, remote sensing, ground- and surface water management, glaciology, and plant ecology, to social sciences, such as environmental psychology and economy, agriculture and tourism.
Within the GLOWA-Danube project - being part of the GLOWA initiative of the German Ministry of Education and Research - we design the Internet based platform DANUBIA as a tool for integrative simulations and support of decision making for a sustainable environmental management in the Upper Danube Basin. Many water-related conflicts arise in this area of almost 80.000 square kilometres concerning, for instance, water quality and water utilisation, environmental protection, flood risks, tourism and vulnerability of mountain environments due to climatic change.
The development of DANUBIA is based on object-oriented software engineering and Web engineering methods and on the Unified Modeling Language UML which is used by all partners as a common graphical notation for modeling the integrative aspects of the system. The crucial concepts behind the design of DANUBIA concern the network and interface structure, the modeling of space and layered simulation areas and the temporal coordination of local models by a global time controller which constitutes the heart of any integrative DANUBIA simulation.
Publications on the above-mentioned and/or related issues are [30], [50], [51], [91], [78], [114], [100], [125].
The architecture of a software system is now widely recognised as the key success factor when it comes to devising, maintaining, and reusing software components, and evolving large scale software systems. One important part of software architecture is the communication between all stakeholders of a software system. Another is to achieve a sound fundament for the software in order to achieve maintainability for the system and re-usability for its parts.
Our research is based on two main strands: the modeling aspect and the technical realisation aspect. Modeling of software systems, especially in the early development phases, involves much effort in communicating with the stakeholders to investigate all requirements and needs the stakeholders have. To communicate results the Unified Modeling Language (UML) is a first choice, as it is the current de facto standard modeling language for software. Integration architectural concepts in the UML meta-model allows us to share the methods and tools for the UML.
The technical realisation aspect is another crucial step to achieve a software that meets all stakeholders' needs and is maintainable and yet extensible. Currently we are extending Java by architectural concepts (Java/A). In this way we achieve a smooth crossover from architectural models to an implementation by providing architectural concepts as fundamental parts in a programming language. Furthermore, Java/A improves the re-usability of software components as it assists programmers in developing them as independent parts.
Lately agile software development processes gained a lot of attention. In contrast to traditional software development processes, these processes promise to provide business value faster than traditional processes and to better cope with changes coming from changing and unclear requirements.
We have successfully applied these processes in the context of the EU project CARUSO, whose objective was the design and implementation of a framework for customised customer relationship management (CRM) applications. Further, we have started to integrate formal methods with these processes.
Performance assessment of software from the early design phase through the implementation has been subject to a great variety of approaches in the past. Performance aspects have to be considered and laid down in certain documents. Thus typical artifacts are created and have to be integrated in the software design as well as in the software development process.
We use UML diagrams to describe all essential aspects of a software and its environment. Dynamic properties of a software are laid down in activity diagrams. The environment that will execute the software is modeled by class diagrams. Finally a simulating software configured by the environment description rates the software model.
Publications on the above-mentioned and/or related issues are [21], [22], [40], [64], [77], [72], [69], [70], [71], [98], and [99].
| Contributing researchers: | Michael Barth, Hubert Baumeister, Rolf Hennicker, Stephan Janisch, Alexander Knapp, Piotr Kosiuczenko (until 2003), Andreas Kraus, Matthias Ludwig, Philipp Meier, Stephan Merz (until 2002), Bernhard Reus (until 2000), Martin Wirsing, Gefei Zhang |
| Cooperations: | ATX Software (Lissabon), DataCall Systems (München), FAST GmbH (München), FJA GmbH (München), ImCon GmbH (München), sd&m Research (München), Remu (Utrecht, Niederlande), Softlab GmbH (München), Soluzione (München), Don Batory (University of Texas at Austin), J. Bendix (Fachbereich Geographie; Phillipps-Universität Marburg), J. Braun (Institut für Wasserbau; Universität Stuttgart), Cristina Cachero (Universidad de Alicante, Spain since 2001), Andrea Corradini (University of Pisa), S. Dabbert (Institut für Landwirtschaftliche Betriebslehre; Universität Hohenheim), Rocco De Nicola (University of Florence), J. Egger (Institut für Meteorologie; LMU), A. Ernst (Wiss. Zentrum für Umweltsystemforschung; Universität Kassel), H. Escher-Vetter (Kommission für Glaziologie der Bayerischen Akademie der Wissenschaften), Jose Fiadeiro (University of Leicester), Stefania Gnesi (Istituto di Scienza e Tecnologie della Informazione "A. Faedo" Pisa), Jaime Gomez (Universidad de Alicante, Spain since 2001), Heinrich Hußmann (LFE Medieninformatik; LMU), H.-B. Kleeberg (IAWG; Institut für Angewandte Wasserwirtschaft und Geoinformatik), H. Kobus (Institut für Wasserbau; Universität Stuttgart), Piotr Kosiuczenko (University of Leicester, since 2003), M. Kuhn (Institut für Meteorologie und Geophysik; Universität Innsbruck), Diego Latella (Istituto di Scienza e Tecnologie della Informazione "A. Faedo" Pisa), Antonia Lopes (University of Lisbon), Wolfram Mauser (Dept. für Geographie und Umweltwissenschaften; LMU), Stephan Merz (INRIA Nancy, since 2002), Ugo Montanari (University of Pisa), Rupert Rockinger(FAST, since 2000), Gustavo Rossi (Universidad de La Plata, Argentinien since 2000), J. Schmude (Institut für Geographie; Universität Regensburg), Wieland Schwinger (Universität Linz, Austria since 2002), R. Sprenger (Ifo-Institut für Wirtschaftsforschung), Andrzej Tarlecki (University of Warsaw), J. Tenhunen (Lehrstuhl für Pflanzenökologie; Universität Bayreuth), Michel Wermelinger (University of Lisbon), J. Wackerbauer (Ifo-Institut für Wirtschaftsforschung), W. Willems (IAWG; Institut für Angewandte Wasserwirtschaft und Geoinformatik), V. Wirth (Institut für Physik der Atmosphäre; Johannes-Gutenberg-Universität Mainz), Julia Zappe (INRIA Nancy, since 2003), |
Constraint Programming is an established paradigm for solving a wide class of problems. Constraint Programming languages allow the programmer to specify a set of relations (constraints) that hold between entities; a constraint solver computes possible values for those entities that satisfy all constraints.
One focus of our work was the declarative specification of constraint solvers. This can be achieved with Constraint Handling Rules (CHR), a programming language based on a committed-choice rewriting system with guarded rules. CHR allow the introduction of user defined constraints into the language. Recent work focused on automatic analysis of properties of CHR programs like termination or complexity. A further area of research was the integration of soft constraint propagation into CHR. We also built several applications using CHR, for example a rent advisor or a program to compute the optimal placement of base stations in wireless indoor communication networks.
Constraint Languages are usually integrated into a logic programming language, giving rise to Constraint Logic Programming languages (CLPs). In recent years there has been interest in combining constraint languages with other programming paradigms to address insufficiencies of CLPs for certain problem domains. We have developed a family of formal calculi (the Constraint Lambda Calculi) that provide a foundation for integrating functional programming languages and constraint solvers, proved confluence properties for the constraint lambda calculi and introduced a denotational semantics. The constraint lambda calculi are parametric in the constraint domain over which they operate and allow the meta-level specification of search and result strategies. They provide most of the facilities of CLPs but can also address problems that are difficult to solve in logic programming languages like the interaction of side effects and constraint solving. Furthermore they allow new advanced programming techniques like hierarchical constraint solver determination. On this foundation we have developed and implemented the ConS/* family of programming languages which extend languages like Scheme, Dylan or Common Lisp with the facilities of the constraint lambda calculi.
Publications on the above-mentioned and/or related issues are [76], [74], [48], [29], [28], [54], [82], [80], [81].
In View-Oriented Programming, we look at a software system from several angles which has the benefit of aiding construction and comprehension of such a system. Each such angle represents an aspect (or view) of a system and is expressed in a formal language (visual or other). This allows one to perform consistency checks between the views or to query them for further information.
We have taken three approaches towards solutions in this area: The first one is to model-check consistency in UML state machines and interactions (see section on model-checking, project Hugo). The second one is algebraic and formalises views as sentences in a language. We allow views to be expressed in different languages and semantics and compare them for consistency using a heterogeneous pull-back construction. The third approach is an extension of Hyperspaces that allows individual methods to be partially distributed into several different Hyper-slices.
Generative Programming is about improving programmer productivity by raising the level of abstraction: Programming usually means translating a specification from a problem domain (which can be expressed in human language or in a notation used by domain experts) to one in the solution domain (usually an executable program expressed in a programming language). Generative programming aids abstraction and communication between domain experts and programmers by letting the latter develop in a notation that is close to the problem domain and by translating it to a language in the solution domain. A related problem is that of Software Product Lines (SPLs): Very often, one needs to produce a family of software systems that are very similar. SPLs is a technique for constructing this family from a common set of base resources.
Work in this area started with a diploma thesis on an editor for a generative programming infrastructure that would later be named AHEAD. AHEAD automatically synthesises members of a product line that can contain a wide array of artifacts (written in human language, domain expert notation, programming language etc.). Furthermore, we developed a formal foundation for AHEAD, Graft to more precisely define its concepts and to analyse software systems developed with it. An analysis can check for semantic and formal correctness and provide information about the system to the programmer.
Publications on the above-mentioned and/or related issues are [94], [108], [61], [102], [101], [121] and [122]
| Contributing researchers: | Thom Frühwirth (until 2001), Matthias Hölzl, Alexander Knapp, Stephan Merz, Bernd Meyer (until 1999), Christopher Rauh, Axel Rauschmayer, Timm Schäfer |
| Cooperations: | Prof. Don Batory (University of Texas at Austin), John Newsome Crossley (CSSE, Monash University), Stephan Merz (INRIA Nancy since 2003), Wolfgang Reif (Universität Augsburg), |
| 1998 | 1999 | 2000 | 2001 | 2002 | 2003 | |
| Full professor(s): | Prof. Wirsing | Prof. Wirsing | Prof. Wirsing | Prof. Wirsing | Prof. Wirsing | Prof. Wirsing |
| Associate professor(s): | Prof. Kröger | Prof. Kröger | Prof. Kröger | Prof. Kröger | Prof. Kröger | Prof. Kröger |
| Research assistant(s): | R. Hennicker, T. Frühwirth, V. Cengarle, S. Merz, B. Reus, A. Knapp | R. Hennicker, T. Frühwirth, V. Cengarle, S. Merz, B. Reus, A. Knapp | R. Hennicker, T. Frühwirth, V. Cengarle,P. Kosiuczenko (since 1.10.00), S. Merz, B. Reus, A. Knapp, | R. Hennicker, T. Frühwirth, P. Kosiuczenko, M. Hölzl, A. Knapp, S. Merz, D. Pattinson, | R. Hennicker, A. Knapp, P. Kosiuczenko, S. Merz, D. Pattinson, M. Hölzl, F. Hacklinger | R. Hennicker, A. Knapp, P. Kosiuczenko (until 30.9.03), D. Pattinson, M. Hölzl, F. Hacklinger, M. Barth (since 1.10.03), Nora Koch (halbe Stelle) |
| Research fellow(s): | H. Baumeister, P. Kosiuczenko, M. Hölzl, N. Koch, A. Kurz, N. Lyabakh, B. Meyer, D. Pattinson, H. Störrle, P. Cenciarelli | H. Baumeister, P. Kosiuczenko, M. Hölzl, N. Koch, A. Kurz, N. Lyabakh, B. Meyer,D. Pattinson, H. Störrle | H. Baumeister, P. Kosiuczenko (until 1.10.00), M. Barth, M. Hölzl, N. Koch, A. Kurz, M. Ludwig, N. Lyabakh, C. Nugraheni, D. Pattinson, J. Rottler, H. Störrle, T. Ströse, J. Zappe | H. Baumeister, M. Barth, F. Hacklinger, N. Koch, A. Kraus, A. Kurz, M. Ludwig, N. Lyabakh, C. Nugraheni, J. Rottler, H. Störrle, T. Ströse, J. Zappe | H. Baumeister, M. Barth, N. Koch, A. Kraus, M. Ludwig, P. Meier, C. Nugraheni, A. Rauschmayr, J. Rottler, J. Zappe, G. Zhang | H. Baumeister, N. Koch (halbe Stelle), M. Ludwig, A. Kraus, P. Meier, C. Nugraheni, A. Rauschmayr, G. Zhang |
Prof. Dr. François Bry, full professor
Prof. Dr. Hans Jürgen Ohlbach, associate professor
| 1998 | 1999 | 2000 | 2001 | 2002 | 2003 | |
| Full professor(s): | 1 | 1 | 1 | 1 | 1 | 1 |
| Associate professor(s) : | 1 | 1 | 1 | 1 | ||
| Research and teaching assistant(s) (*) | 4 | 4 | 4 | 4 | 4 | 4 |
| Research assistant(s) (without teaching duties) (**) | 2 | 2 | 2 | 2 | 1 | 3 |
(*) "Wissenschaftlicher Mitarbeiter auf Landesstelle mit
Lehrverpflichtung"
(**) "Wissenschaftlicher Mitarbeiter aud Drittmittelstelle oder
mit Promotions- oder Postdoc-Stipendium ohne Lehrverpflichtung"
A deductive Web query language, called Xcerpt, is under development [49] [52] [58] [59]. Xcerpt is based on a novel form of unification, called "simulation unification" [57]. A first prototype has been completed and is used with an original visual programming interface [59] [63]. Elements for a formalisation of the declarative semantics of Xcerpt are given in [62]. The procedural semantics and the run time system of Xcerpt are based on constraint reasoning [57]. Temporal types and operations for Xcerpt are discussed in [60]. First experiments with Xcerpt point to the conciseness and simplicity of Xcerpt programs compared to e.g. XQuery programs and suggest that Xcerpt might considerably ease the implementation of adaptive Web systems.
The novel notion "complete answer aggregates" of aggregated answer to queries against XML and/or semistructured data is described in [3] and [35]. The notion aims at sharing subparts common to the various answers to a same query. A prototype system has been developed that computes complete answer aggregates to queries expressed in an XPath-like language and using which a user can "browse" aggregated answers. The prototype, its application to test beds, and its visual user interface are described in [55].
A method called SPEX for evaluating continuous XPath queries against XML data streams is presented in [69]. The approach first translates general XPath queries into equivalent XPath queries without so-called reverse axes such as "parent" or "ancestor" [47]. Then, such XPath expressions are translated into networks of push-down automata. The method has a polynomial combined complexity. A prototype and its evaluation on bench marks are reported about in [53].
Building upon the Data Guide approach to indexing the structure of XML or semistructured data, a method called "Content Aware Data Guide (CADG)" for an indexing of both, structure and textual content has been developed. The method, its implementation, and its performances on bench marks are described in [68].
Further publications on the above-mentioned and/or related issues are [31] [40] [42].
| Contributing researchers: | François Bry, Norbert Eisinger, Hans Jürgen Ohlbach, Dan Olteanu, Sebastian Schaffert, Stephanie Spranger, and Sacha Berger. |
| Co-operations: | Klaus U. Schulz (Centre for Information and Language Processing, University of Munich) and Holger Meuss (European Southern Observatory). |
| Research Grants: | Two full time positions (Stephanie Spranger since January 2003 and Paula Pàtrânjan since November 2003) funded by the DFG within the doctoral programme "Graduiertenkolleg Logik in der Informatik" |
Logic provides with a convenient basis for declarative languages such as Xcerpt. However, pure logics provide little means for modeling real world notions and objects. In project WebTNSS a logic is developed that has such means. The focus is on the modeling of "common sense" time and calendric and location notions [66]. These notions are becoming more and more important on the Web with the emergence of mobile Web applications. A related project is devoted to integrating elementary arithmetics in description logics and ontology languages [36]. The goal of both projects is to enrich logics and logic-based languages with numerical aspects of sets such as price, age, duration, etc.
The implementation of adaptive Web systems using Web query languages is investigated [48] [56] [67]. First investigations suggest that the approach should considerably simplify the implementation of adaptive Web systems.
| Contributing researchers: | François Bry, Bernhard Lorenz, Hans Jürgen Ohlbach, Stephanie Spranger |
| Co-operations: | Michael Kraus (Keio University, Tokyo) and Kazuhiro Kitagawa (W3C Kyoto). |
| Research Grant: | 55 000 Euro from the Max Plank Institute for running expenses in 2000 |
The analysis, integration, optimisation, and automatic generation of rule-based constraint solvers has been investigated focusing on constraint solvers implemented in the rule-based language CHR. Methods for the analysis and optimisation of constraint solvers are presented in [1] [24]. Methods for an automatic generation of constraint solvers from complete, incomplete, and intentional specifications are described in [9] [28] [41] [45] [46] [54] [70] [65]. The integration of constraint solvers is addressed in [64].
Related results and constraint programming tools tools are presented in [8] [23] [29] [37] [43] [51].
| Contributing researchers: | Slim Abdennadher and Michael Marte. |
| Co-operations: | Henning Christiansen (Roskilde University), Thom Frühwirth (University of Ulm), and Christophe Rigotti (ENSA Lyon, France) |
| Research Grant: | One full time position (Michael Marte from February 1999 to
January 2002) funded by the DFG
within the doctoral programme "Graduiertenkolleg SIL
(Sprache, Information, Logik)" 14 000 Euro for travel from the BFHZ-CCUFB in 1999 und 2000 |
Various applications of logic, e.g. type theory, to programming and software maintenance are investigated in [34] [38].
A formalisation of a (refinement of the) stable model theory in terms of a is given in [39]. The salient aspect of this formalisation is that it is expressed in the manner of classical model theory in terms of a truth-function function defined recursively on the structure of formulas. This definition makes it possible to evaluates logic programs according to the (refined) stable model semantics using standard proof methods such as resolution.
Several contributions to the theorem proving framework introduced with the SATCHMO prover have been investigated up till 1998. A deduction method complete for both refutation and finite satisfiability is presented in [2] [10]. Such a method is useful for artificial intelligence applications such a processing natural and/or controlled language specifications [25]. The generation of minimal models, an issue of importance for artificial intelligence applications such as planning, is addressed in the framework of the SATCHMO theorem prover in [30]. The application of a model generation prover to problem solving and to the specification of database integrity constraints is addressed in [11] [13] [19]. Implementations of provers are described in [12] [15]
Further reports on related issues are [20].
| Contributing researchers: | François Bry, Norbert Eisinger, and Reinhard Kahle. |
| Co-operations: | Peter Baumgartner and Ulrich Furbach (both University of Coblence, Germany), and Adnan Yahya (Bir Zeit University, Palestine). |
| Research Grants: | Three full time positions (Sunna Torge from November 1995 to August 1999, Thomas Brüggemann from December 1995 to April 1998 and Reinhard Kahle from 2000 to 2001) funded by the DFG within the doctoral programme "Graduiertenkolleg SIL (Sprache, Information, Logik)" and one full time position (Heribert Schütz from November 1996 to April 1999) funded by the Bavarian Ministery for Education and Research |
Timetabling issues have been investigated in various contexts, e.g. high school, universities, and hospital. Timetabling systems have been developed using constraint logic programming and extensively tested on benchmarks [6] [7] [17] [18] [22] [33] [5] [26]. Constraint logic programming has been applied for implementing a 'rent advisor' according to German city office regulations fixing upper bounds to apartment rents. Constraint programming makes it very easy to reason with interval and/or missing data instead of exact data [14] [21] [27]. Knowledge assimilation in deductive databases has been investigated in [16]. The interactive analysis of high-dimensional association structures has been investigated in [32]. An overview of current Bioinformatics databases is given in [50]. An approach to Web-based co-operative work has been specified, implemented, and tested. Web services for teaching are addressed in [61]. Proposals for an enhanced modeling of electronic books are made in [44].
| Contributing researchers: | François Bry, Norbert Eisinger, Hans Jürgen Ohlbach, Sunna Torge. |
| Co-operations: | Hendrick Decker (Siemens, Munich, Germany, and Instituto Tecnológico de Informática, Valencia, Spain) and Iris Pigeot (Institute for Statistics, University of Munich). |
In addition, members of the Unit have contributed to several additional conferences as program committee members or as reviewers.
Slim Abdennadher has a full professorship (since September 2003) at the German University in Cairo, Egypt.
Two former members of the Unit, Tim Geisler and Dr. Heribert Schütz, have founded in 2000 a start-up company, webXcerpt, developping Software and offering services in the areas of Web data retrieval and querying.
Between 1998 and 2003, external funding has been acquired by the unit for five full-time PhD Students in the framework of PhD programs (or "Graduiertenkollegs") of the German Foundation for Research "DFG".
In 2001 and 2003 an application for a Network of Excellence "REWERSE" on "Reasoning on the Web" in the 6th Framework Programme of the EU Commission has been successfully prepared under the co-ordination of François Bry. REWERSE will start in 2004 and last over 4 years, involves about 100 researchers from about 30 institutions, and has an overall budget of 5,5 Million Euro of which 1,6 Millions are assigned to the Unit "Programming and Modelling Languages".
In 1999 and 2000 the unit has received 14000 Euro from the BFHZ-CCUFB (Bayerisch-Französisch Hochschulzentrum - Centre de Coopération Universitaire Franco-Bavarois) for joint research activities with colleagues from France.
Prof. Dr. Martin Hofmann, full professor
| 1998 | 1999 | 2000 | 2001 | 2002 | 2003 | |
| Full professor(s): | 1 | 1 | 1 | 1 | 1 | 1 |
| Associate professor(s) : | ||||||
| Research and teaching assistant(s) (*) | 3 | 3 | 4 | 4 | 3 | 3 |
| Research assistant(s) (without teaching duties) (**) | 2 | 4 | 5 | 6 |
(*) "Wissenschaftlicher Mitarbeiter auf Landesstelle mit
Lehrverpflichtung"
(**) "Wissenschaftlicher Mitarbeiter aud Drittmittelstelle oder
mit Promotions- oder Postdoc-Stipendium ohne Lehrverpflichtung"
This report covers the period from September 2001 because the chair holder, Professor Martin Hofmann took up his position only then. Publications of those current members who have been employed before September 2001 are included as are publications of Martin Hofmann in case they are of importance for the current activities of the unit.
The former chair Peter Clote is now full professor of biology at Boston College, the former assistant Thorsten Altenkirch is lecturer in computer science at the University of Nottingham, and the former assistant Rolf Backofen is full professor for bioinformatics at the University of Jena.
Prizes: Martin Lange received the Best Student Paper Award at the 13th Int. Conference on Concurrency Theory in Brno, Czech Republic, for his paper "Local Model Checking Games for Fixed Point Logic with Chop" (2002). The program committee of the int. conference "Typed Lambda Calculi and Applications 2001" in Kraków, Poland, chaired by Samson Abramsky (University of Oxford), awarded Ralph Matthes with the prize for the "most promising paper, TLCA 2001".
Apart from the research projects below, there are two important project involvements:
| Contributing Researchers: | Martin Hofmann (contact person), Steffen Jost, Hans Wolfgang Loidl, Olha Shkaravska |
| Cooperations: | Cooperation with University of Edinburgh (project partners) |
| Research Grant: | Funding as project IST-2001-33149 (European Commission's Fifth Framework Programme) |
The project members are developing the foundations for guaranteeing bounded resource consumption of mobile code. The goal of this effort is to enhance the security of large-scale distributed systems by preventing third-party code from exhausting available resources. A proof-carrying-code infrastructure is used, where resource certificates are shipped together with mobile code. Key techniques used in this project are automated theorem proving (currently using Isabelle), for validating the resource certificates, and high-level type-systems [52,10,11], which extend classical types for functional languages with information on resource consumption.
| Contributing Researchers: | Jan Johannsen |
| Cooperations: | with University of Toronto (Alasdair Urquhart, Toniann Pitassi), Universitat Polytecnica de Catalunya (Albert Atserias, Maria Luisa Bonet, Juan Luis Esteban, Nicola Galesi) |
| Research Grant: | Funding as Independent Junior Research Group in the DFG Emmy-Noether-Programme, under grant no. Jo 291/2-1 and Jo 291/2-2, 10/1999-6/2003, employed: N.S. Narayanaswamy, Ulrich Rührmair |
In this project the complexity of proof systems for classical propositional logic is studied. In particular we investigate the minimal lengths of proofs in refinements and extensions of the Resolution calculus. Lower bounds on the proof lengths in these systems imply runtime lower bounds for certain types of backtracking algorithms, e.g. for satisfiability testing. Another goal is to study the complexity of monotone boolean circuits, which can be a useful tool for the main project objective above. Publications: [50,39,19,20]
| Contributing Researchers: | Martin Hofmann (contact person), Jan Johannsen |
| Cooperations: | with LMU Department of Mathematics (Helmut Schwichtenberg, Klaus Aehlig), Universidade Nove de Lisboa (Isabel Oitavem), University of Ottawa (Phil Scott), Brandeis University (Harry Mairson), INRIA Nancy (Jean-Yves Marion) |
| Research Grant: | Partially funded by the EU working group Applied Semantics II, coordinated by Martin Hofmann. |
The goal of this project is to find characterisations of computational complexity classes that do not rely on a specific model of computation and explicit resource bounds, but on more conceptual means, e.g. characterisations by type systems, programming languages or logical systems. Apart from insights into structure of complexity classes this has applications to resource certification of programs [10] and code optimisation, in particular of space usage [52]. Publications: [38,9,52,21,10,11,1]
| Contributing Researchers: | Andreas Abel, Ralph Matthes (contact person) |
| Cooperations: | with Tarmo Uustalu, Senior Researcher at Tallinn Technical University, Estonia |
| Research Grant: | Partially funded by the EU working group "TYPES", project no. 29001 in IST programme, and by the "Graduiertenkolleg Logik in der Informatik" of TU and LMU. |
Nested datatypes, as introduced by Bird (Oxford) and his coworkers, can be understood semantically as initial algebras of endofunctors on a functor category. In this framework, notions of substitution for very general data structures can be represented [15]. The project is mainly concerned with a programming language approach: For higher-order typed lambda calculi, schemes of iteration and primitive recursion for inductively defined families of datatypes are identified and the programs expressible in these formalisms proven terminating, also for infinite datatypes. Typical examples are representations of terms involving binding such as lambda abstraction and explicit substitution operators. References: [42,13,14,15]
| Contributing Researchers: | Jan Johannsen, Martin Lange |
Verification of hardware or software systems is often done using logics that are interpreted over program behaviours, i.e. modal, temporal or dynamic logics, etc. Even though most of them are known since the '70s or even earlier, there are still interesting fragments or variants of the standard logics with decision problems whose exact complexity bounds are still unknown. The aim of this research is to determine these bounds. Reference: [8]
| Contributing Researchers: | Hans-Wolfgang Loidl |
| Research Grant: | Funding as project ARC D/03/20257 of the DAAD with Philipps-University Marburg, Heriot-University Edinburgh, and St Andrews University. |
The goal of parallel functional programming is to speed-up computations in functional languages by executing them on several processors in parallel. In contrast to conventional parallel languages, no detailed specification of the coordination between the parallel processes has to be given. It is only necessary to annotate those expressions in the program that might be evaluated in parallel, and a sophisticated runtime environment automatically manages the coordination and distribution of work and data. We have developed a parallel extension of Haskell98, called GpH, and are studying program analyses and program development in this high-level parallel language [22], the efficient implementation of the runtime-environment on a range of parallel architectures [12,23] and extensions for mobile code [24].
| Contributing Researchers: | Favio Miranda |
| Research Grant: | PhD project funded by DAAD and CONACYT (Mexico's National Science and Technology Council) |
This research is a chapter of the synthesis problem. In our case the specifications are given by formulas of a logical system of (co)inductive definitions which is an extension of second order logic with equality (AF2) whereas the programming language is a strongly normalising system of (co)inductive types which is an extension of the second order polymorphic lambda calculus. Using a realisability interpretation we can automatically transform the proof of the specification into a proof of the fact that the program realizes it and therefore additional program verification methods are not needed. The new contribution consists in extending the method to specifications that include inductive definitions like lists and trees and even co-inductive definitions like streams or infinite trees. The paper [26] reports early stages of this research.
| Contributing Researchers: | Hans Dietmar Jäger |
| Research Grant: | PhD project funded by "Bayerisches Staatsministerium für Wissenschaft, Forschung und Kunst" within the SIGNAL project, see above |
Today learning is rather a life-long process than a one-time task. In this regard media and PC competence is a key qualification. Keeping that in mind this project is about relating prerequisites of learners to their learning process in distance learning postgraduate professional education. We centre our main focus on media and PC competences in a virtual learning environment with extremely heterogeneous participants. Future work will deal with setting up a virtual learning environment with specific functions. A didactical, pedagogical and social concept will be made to support learners in a cooperative learning process. Main goal of this project is to observe the influence of deficits in handling new media and computers on learning progress and to map out a strategy to countervail this behaviour. Reference: [7]
| Contributing Researchers: | Martin Lange |
| Cooperations: | Humboldt-University Berlin |
Most logics used for verification purposes turn out to be embeddable into the Modal mu-Calculus, a modal fixpoint logic that is often said to be very expressive. However, due to a very important result by Janin/ Walukiewicz - Modal mu-Calculus is equi-expressive to the bisimulation invariant fragment of Monadic Second Order Logic - its expressive power is rather limited. All properties defined by formulas of the Modal mu-Calculus are regular in the language theoretic sense. Thus, it is not even possible to express that something behaves like a bounded stack. This research aims at logics beyond the Modal mu-Calculus and suitable model checking algorithms for them. Reference: [25]
| Contributing Researchers: | Ralph Matthes |
| Cooperations: | with LMU Department of Mathematics (Helmut Schwichtenberg) |
Typed lambda calculi are the notation systems of choice for proofs in intuitionistic natural deduction. Since classical logic can be embedded into intuitionistic logic, also the former can be studied by the operational techniques available for lambda calculi. The operational reading of classical logic is of deep interest in programming language theory as an idealised functional programming language with control. This project aims at novel termination proofs for programs in these language fragments (such as Parigot's lambda-mu-calculus) and tries to incorporate complex datatypes. An alternative technique to CPS translations has been discovered. References: [43, 5]
Prof. Dr. Heinrich Hußmann, full professor
| 1998 | 1999 | 2000 | 2001 | 2002 | 2003 | |
| Full professor(s): | 1 | |||||
| Associate professor(s) : | ||||||
| Research and teaching assistant(s) (*) | 4 | |||||
| Research assistant(s) (without teaching duties) (**) | 1 |
(*) "Wissenschaftlicher Mitarbeiter auf Landesstelle mit
Lehrverpflichtung"
(**) "Wissenschaftlicher Mitarbeiter aud Drittmittelstelle oder
mit Promotions- oder Postdoc-Stipendium ohne Lehrverpflichtung"
In the project Media of the Future we investigate the use of new types of media and devices for presentation and interaction with information. In the course of the project we investigate commercially available products as well as technologies available in the research community and assess their applicability for specific application domains. The focus is on input and output devices for virtual and augmented reality, technologies for mobile multimedia applications, and physical user interface. Beyond that we investigate technologies and methods that can be used for implicit user interfaces; of particular interest are sensors for capture and ambient media for peripheral information provision. The evaluation of technologies lead to a laboratory setup where specific technologies can be tested and assessed in detail with regard to specific application domains and user's needs. Examples of technologies that are available in the lab are an interactive whiteboard (Smart-Board), a 3D head-mounted display unit, data gloves for input, and prototypes of physical user interfaces (Smart-Its). This is complemented by software and development tools for these systems.
Contributing researchers: Heinrich Hußmann, Albrecht Schmidt, Arnd Vitzthum
Multimedia applications are widely used for presentations, learning systems, and teaching tools. Currently development support and authoring tools are tailored to very specific application domains. The development of complex interactive multimedia applications is inadequate. Model driven development, reusable components, and interface abstractions - standard practice in software engineering - are poorly supported. In this project it is investigated how methods and tools successfully used in software engineering can be applied to improve the development process of complex multimedia applications. We are interested how such applications are developed, especially the cooperation of designer and software engineers is investigated. Existing development system for conventional software and authoring tools for multimedia applications have been studied and compared. These results are the basis for the development of new tools and methods for modelling and creating of multimedia applications.
Contributing researchers: Heinrich Hußmann, Andreas Pleuß
In more and more domains the production of audiovisual media is done entirely based on digital technology. Images, Sounds, and Video are captured, recorded, communicated, and stored in digital form. Additional processing steps are also carried out digitally. The approach changes work processes and allows new forms of media production. In the project digital media production a concept for a research laboratory was created and currently the installation is on its way. The lab allows us to do investigate all steps that are involved in the process of digital media production. We investigate processes where video, animation, and 3D graphics merge. A further topic is digital audio production and distribution channels. We are especially interested in new models and data formats for digital radio and in particular Internet radio. Based on these technologies we investigate new approach and technologies for digital media production.
Contributing researchers: Heinrich Hußmann, Albrecht Schmidt, Siegfried Wagner
The entire project is third party funded by the federal ministry of education and research (BMBF). Intermedia is an interdisciplinary project at the LMU and is part of the BMBF's program "Internet economy". Within Intermedia six research groups cooperate, namely Hußmann and Linnhoff from Computer Science, Picot and Hess from Business Studies, Brosius and Wirth from Communication Science. The overall objective of sub-project 4 - Intermedia - is model-based application development for information intermediaries (Modellbasierte Anwendungs-entwicklung für Informationsintermediäre). The aims of the project are the development of description language for business processes and supporting tools and infrastructure. The description language will be based on UML (Unified Modeling Language) and is designed to be appropriate for modelling complex business process on an abstract level. The language will be the basis for analysing business processes with regard to economic aspects. In the course of the project a prototypical reference implementation of an application that processes the description language will be designed and implemented. The main function of this application is to transform descriptions of business processes into executable programs. These generated programs will be web applications using HTML as mechanism for creating the user interface.
Contributing researchers: Heinrich Hußmann, Richard Atterer
Prof. Dr. Ralf Zimmer, Full Professor
Prof. Dr. Volker Heun, Associate Professor
| 1998 | 1999 | 2000 | 2001 | 2002 | 2003 | |
| Full professor(s): | 1 | 1 | 1 | |||
| Associate professor(s) : | 1 | |||||
| Research and teaching assistant(s) (*) | 1 | 5 | 6 | |||
| Research assistant(s) (without teaching duties) (**) | 3 |
(*) "Wissenschaftlicher Mitarbeiter auf Landesstelle mit
Lehrverpflichtung"
(**) "Wissenschaftlicher Mitarbeiter aud Drittmittelstelle oder
mit Promotions- oder Postdoc-Stipendium ohne Lehrverpflichtung"
The unit "Practical Computer Science and Bioinformatics" started in the winter term 2001/2002 with Prof. Zimmer and one PhD student (Florian Sohler), Prof. Heun started in April 2003. Currently, the unit consists of three postdocs and six PhD students (plus additional three PhD students employed at the Fraunhofer Institute SCAI, Sankt Augustin). The activities concentrate on the establishment of the Bioinformatics Initiative Munich (BIM) and algorithmic aspects of bioinformatics. In computer science the unit focuses on areas relevant for bioinformatics applications:
Currently, work in this area is organised in two projects: DFG project PROSEQO and PSY. The goal of the DFG funded PROSEQO (Protein sequence structure optimisation, 2 years, 2 full time scientists and 2 students) within the DFG focus program "Verfahren zur Analyse großer genomischer Datenmengen" is the development of fast, accurate and flexible algorithms and software for protein structure prediction, i.e. the computation of the three dimensional structure models for genomic or amino acid sequences ("The second half of the genetic code"). The problems involved are hard from an algorithmic viewpoint (NP complete) and difficult with respect to modeling aspects (protein folding).
The project PROSEQO develops and implements heuristic combinatorial optimisation methods for protein structure prediction, determines good and efficiently computable bounds for structure prediction, exploits appropriate biological constraints, defines a formal language for the specification of such constraints (e.g. using a XML based markup language) and uses it in optimisation methods. The project constructs feature based views of the protein sequence structure space, works on multi-criteria clustering of sequences and structures at the same time and develops methods for classification and alignment of proteins based on various criteria and scoring functions. The effective and discriminative construction and evaluation of those criteria and scoring systems are also done in the project.
The project PSY (protein structure analysis) aims at integrating advanced students and student staff into a collaborative effort to implement a competitive protein structure prediction server. This involves the development of new methods and derived data resources, the exploitation of available programs, databases and web services, the intelligent combination of these methods, and the effective use of various supercomputer and workstation cluster computer facilities.
Related publications: [4, 5, 6, 7, 13, 15, 21, 22, 26, 27, 29, 30, 35, 36, 37, 52, 55, 56, 58, 60, 63, 64, 65, 66, 68, 69, 74, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 97, 99, 102, 104, 105, 106, 107]
| Contributing researchers: | Dipl.-Inform. Alessandro Macri (since 08/2002), Dipl.-Inform. Jan Erik Gewehr (since 04/2003), Ralf Zimmer. |
| Collaborations and external PhD students: | FhI SCAI, Sankt Augustin, BioSolveIT GmbH, Sankt Augustin. Dipl.-Inform. Heinz-Theodor Mevissen (FhI), Dipl.-Inform. Daniel Hanisch (FhI), Dipl.-Math. Niklas von Öhsen (FhI). Dr. Ingolf Sommer (MPI Informatik, Saarbrücken). |
| Research grant: | DFG, two years, two full time researcher positions, two student positions |
| Thesis work: | Andreas Spitzmüller (ongoing) |
Within several projects funded by Aventis (BEX, two years, 2 full time scientists) and by the BMBF (BOA, 3 years, 2 scientists) new innovative methods for the analysis of expression data and of metabolic and regulatory networks are being developed. High throughput data obtained via DNA chips allow the genome wide measurement of cellular activities at the molecular level and thus, the comparison of different cell types, cell states and of normal/diseased tissue. In particular, graph and Petri net models and algorithms combine network analysis and statistical evaluation for the combined interpretation of expression data. Goal is the construction of disease models, the identification of new drug target molecules as starting points for new drug candidates, therapies and diagnostic means. Methods for mining metabolic and regulatory relationships from scientific texts and data extraction from appropriate databases are developed and used to generate large networks building the basis of new hypotheses concerning diseases and molecular mechanisms. The methods are applied for experimental data and measurements from several disease groups of the pharmaceutical company Aventis (Frankfurt). In addition, methods to design and equip new taylor-made DNA-chips.
The project BEX concentrates in this area on the development and implementation of new algorithms and tools. The project BOA focuses on the application and tuning of the methods as well as the development of new concepts for expression data analysis, in particular for osteoarthritis research. BOA provides bioinformatics analysis tools for and within the BMBF molecular medicine Leitprojekt "Therapie und Diagnose der Osteoarthrose" in collaboration with about 20 partners from pharmaceutical industry, Biotech companies, start-ups, research institutes and university hospitals for the analysis of expression data and networks of osteoarthritis. In the BEX and BOA projects two successful software tools have been developed: ProMiner and ToPNet. ProMiner allows for a sensitive and accurate search for gene and protein names in large scientific text bodies (such as the about 15 millions of PubMed abstracts) and to construct simple co-occurrence networks from the searches. ToPNet allows for visualisation and analysis of networks of various types, e.g. text mining co-occurrence networks, interaction networks derived from databases such as DIP or Swissprot, or detailed reaction or regulation networks as listed in specialised biochemical databases such as Transpath. In contains a broad range of interactive selection, visualisation, searching and analysis tools for such networks and associated annotations. Theses annotations can be all kinds of functional data, mappings to classifications and ontologies, or measurement data such as DNA-chip expression data or proteomics measurements. ToPNet uses a set of mapping files that facilitates the connections of the various network nodes, genes and proteins, and the annotations and measurements. A couple of new algorithms have been developed, published and integrated into ToPNet, which allow to identify interesting genes/proteins and pathways/subnetworks based on the network structure and the respective experimental data. These methods include pathway scoring, co-clustering (of network and expression data), significant area search, pathway queries, and clustering (of expression data). Thus, ToPNet allows for efficient and interactive interpretation of experimental data based of automatically and/or semi-automatically assembled contexts of biological knowledge.
Related publications: [1, 2, 3, 14, 23, 24, 25, 31, 32, 33, 34, 54, 57, 59, 74, 75, 76, 92, 93, 94, 95, 96, 97, 98, 103]
| Contributing researchers: | Dr. Roger Abseher (04/2002 - 04/2003), Dipl.-Biotech. Katrin Fundel (since 04/2002), Dr. Robert Küffner (since 09/2003), M.Sc. (computer science), Florian Sohler (since 10/2001), Ralf Zimmer. |
| Collaborations and external PhD students: | Dipl.-Inform. Daniel Hanisch and Dipl.-Inform. Heinz-Theodor Mevissen, FhI SCAI (Sankt Augustin); Dr. Sabine Trochim, BioSolveIT GmbH (Sankt Augustin); Bioinformatics group Aventis Pharma GmbH (Frankfurt); Disease group DG Thrombosis & Degenerative Joint Diseases, Aventis Pharma GmbH (Frankfurt), GPC Biotech (Martinsried); BioSolveIT GmbH (Sankt Augustin); Fraunhofer Institut SCAI (Sankt Augustin); GSF München; BioTez Berlin; Invitek Berlin; University Erlangen; University Münster; University Mainz; University Frankfurt; Hospital Rummelsberg. |
| Research grant: | Aventis (BEX), Aventis+BMBF+FhI (BOA): three years, two (+ two) research positions, four student positions, Server (58 TEUR), software and travel funds. |
| Thesis work: | Cornelia Donner (8/2003 [14]), Caroline Friedel (8/2003 [23]), William Holton (11/2003 [54]), Simon Tietze (ongoing) |
The BMBF funded project ProBio/MAMS develops methods and bioinformatics tools for the construction and analysis of protein chips in co-operation with several universities, research institutes, as well as companies. The protein chips developed in the project are based on a innovative detection method for bound protein molecules via mass spectrometry and new binding mechanisms for these proteins via chip-attached RNA molecules (so called aptamers). The new chip will be applied to data from the application areas blood coagulation (U. Bonn) and viral infections (U. Cologne) for the identification of new targets and the validation of target candidates. The goal of the LMU project is the development of methods for the joint and combined analysis of networks, mRNA expression and protein expression data.
The PROBIO project is part of a larger BMBF funded project for the development of an aptamer based protein chip. Aptamers are short nucleic acid sequences that have been selected for their specific binding to a given (protein) target. The aim of PROBIO is to provide the bioinformatics support for finding suitable candidate target and perform the analysis of the results. A number of different research topics are being addressed within the scope of this project:
Related publications: [8, 9, 10, 11, 12, 17, 18, 19, 20, 24, 25, 28, 32, 33, 51, 52, 53, 59, 60, 62, 67, 73, 76, 79, 80, 87, 88, 89, 97, 98, 99, 100, 101, 103]
| Contributing researchers: | Dr. Joannis Apostolakis, Dr. Christoph Best, Dipl.-Biotech. Katrin Fundel, Ralf Zimmer |
| Collaborations and external PhD students: | Dipl.-Math. Niklas von Öhsen (FhI SCAI); Stiftung caesar, Bonn; FhI SCAI, Sankt Augustin; University Bonn, Institute for experimental hematology and transfusion medicine; University Cologne, Institute of Virology; Qiagen GmbH, Hilden; NascaCell AG, Tutzing; Bruker Daltronic GmbH, Bremen. |
| Research grant: | BMBF, FhI: three years, two (+ one) research positions, two student positions, Server (58 TEUR), software and travel funds. |
| Thesis work: | Thomas Obkircher (11/2003 [72]), Daniel Güttler (ongoing) |
Cheminformatics deals with theoretical models and computer science methods and to address problems and data obtained in chemistry applications. One problem is the prediction of binding free energies and optimisation of energy/scoring functions for the prediction of molecular structures and properties [11, 12, 67, 73]. A second interest is in developing machine learning methods for classification, regression and feature extraction [10, 51]. A third problem is the understanding and simulation of the dynamics of molecular and cellular processes. Work done on this topic involves the dynamics of Vancomycin, the analysis of co-operativity effects in binding and on protonation states. A major problem concerns protein folding: research is on the development and application of methods for calculating folding trajectories, conformational transitions and predicting protein structures of peptides and small proteins. [8, 9, 17, 18, 19, 20]
The DFG funded Template project (co-operation with MPI Saarbrücken and University of Düsseldorf) has as its main aim the development of methods for artificial receptor design. Artificial receptors are relatively low molecular weight molecules that nevertheless can bind a specific ligand with high affinity and specificity.
Future efforts will be focused on a simple theoretical framework for the analysis of gene expression and cellular gene expression dynamics. Chem- and bioinformatics methods and models will be combined to obtain novel insights into biological processes and find possible applications in the field of drug design and biotechnology.
Related publications: [8, 9, 10, 11, 12, 17, 18, 19, 20, 28, 51, 62, 67, 71, 73, 97]
| Contributing researchers: | Dr. Joannis Apostolakis, |
| Collaborations and external PhD students: | Dr. Philippe Ferrara, NIH; Dr. Andreas Kämper, MPI Saarbrücken; Prof. Dr. Christel Marian, University of Düsseldorf. |
| Research grant: | DFG, template project: two years, one researcher position |
The prediction of the spatial structure of a protein is one of the most important open problems in molecular biology. For the standard combinatorial model (the so-called HP model proposed by Kenneth Dill), only a few simple approximation algorithms are known, which achieve poor approximation ratios and produce rather artificial conformations. To overcome the artificial predictions, different and more appropriate relaxed discretisation of the 3-dimensional space as well as the relaxation of the embedding constraints have to be studied.
In the 3-dimensional extended cubic lattice we proposed a general folding algorithm which achieves for all protein sequences an approximation ratio of 59/70. A more sophisticated folding algorithm which can be applied to a restricted subset of HP-sequences yields an approximation ratio of 37/42. Although it is difficult to compare the approximation ratios for protein structure prediction algorithms on different lattice models, it should be mentioned that this is the best known approximation ratio for such algorithms. On the other hand, this is the first time that folding algorithms for a `natural' subclass of HP-sequences have been investigated. A strong indication that the considered subclass of HP-sequences is a `natural choice' is the fact that more than 99.5% of all sequences in protein data base SWISS-PROT belong to the considered subclass. Finally, the running time of both approximation algorithms are linear.
Related Publications: [38, 39, 44, 45, 47, 48, 49, 50]
| Contributing researchers: | Volker Heun, Dipl.-Inf. Johannes Fischer (since 10/03), Dipl.-Inf. Simon W. Ginzinger, M.Sc. (since 10/03) |
| Research grant: | DFG (BIM) 2+3 years. |
One step in the analysis of gene expression data is the clustering of co-expressed genes. Due to the large amount of generated data and due to the fact that the data are perturbed by noise, rigid mathematical models for the analysis as well as efficient algorithms are necessary to overcome the involved problems. We presented a sound mathematical model dealing with the large error rates and proposed two algorithms (one based on probabilistic methods, the other based on spectral graph theory) to solve the involved clustering problem efficiently. These algorithms recognise perturbed cluster structures assuming that the noise is identically and independently distributed. The algorithms are completely analysed and experimental results on synthetic data as well as on real data demonstrate their capability to solve the clustering problem efficiently.
Related publications: [16]
| Contributing researchers: | Volker Heun. |
| Collaborations: | Dr. Jens Ernst, Prof. Dr. Ernst W. Mayr, Dr. Ulrich Voll (Technische Universität München). |
An important algorithmic problem is to develop efficient genome comparison methods to find the minimum number of global mutation operations to transform one genome sequence into another and thereby characterising the evolutionary relationship between the corresponding organisms. This approach guarantees with a high degree of certainty that existing evolutionary relationships will be discovered.
This combinatorial problem is of particular interest to theoretical computer science because it has been shown to be NP-complete and therefore, in full generality, not solvable efficiently due its high inherent complexity. On the one hand this motivates the search for efficient approximation algorithms guaranteeing high quality results and short running times in real life experimental settings. And on the other hand it justifies attempts to achieve a reduction of complexity through modifications to the underlying formal model making the biological problem accessible to computer science.
We study the complexity of different versions of the genome rearrangement problem and evaluate existing formal models in terms of their computational feasibility - most of the questions arising in this context are currently open. In the same context we aim at designing new and realistic models which should still allow algorithmic solutions. The main focus, however, is on developing efficient combinatorial methods yielding either exact or approximate results with a high guaranteed approximation quality.
| Contributing researchers: | Volker Heun, Dipl.-Inf. Johannes Fischer (since 10/03), Dipl.-Inf. Simon W. Ginzinger, M.Sc. (since 10/03) |
| Research grant: | DFG (BIM) 2+3 years, |
Visualisation of biological networks (e.g., metabolic or regulatory networks) is important for their better understanding. Such networks will usually be modeled with graph-theoretic concepts like Petri nets or attributed (hyper-)graphs. Due to the large amount of data, it is not useful to generate a single map representing it. Hence, algorithmic methods are needed for organising data in different layers representing different levels of abstraction. Therefore, on the one hand, methods are sought for structuring and partitioning graphs and their underlying data, and on the other hand, methods to reveal similarities within the same layer of abstraction as well as between different levels.
Provided suitable computational models of biochemical networks, our goal is to develop efficient algorithms to detect similarities for structuring given graphs as well as to find close relationships between parts of the given graph. Although this problem is in general NP-hart (subgraph isomorphism), the additional biological information stored as attributes at nodes and (hyper-)edges might be helpful to design efficient algorithms. On the other hand, specific models for graph similarity are required (like edit-distances) to quantify the similarity of graphs. Such models are useful if the induced similarity measures can be determined efficiently.
Another aspect is the investigation of fluxes or pathways within biochemical networks. Here we are interested in the decomposition of a given network or fluxes into fundamental ones under the constraints of involved products or enzymes. Such algorithms are required as a first step of analysing the function of a given biochemical network and to find possible alternative pathways of a given flux or to inhibit certain pathways.
| Contributing researchers: | Volker Heun, Dipl.-Inf. Johannes Fischer (since 10/03), Dipl.-Inf. Simon W. Ginzinger, M.Sc. (since 10/03) |
| Collaborations: | Dipl.-Biol. Arno Buchner, Dr. Jens Ernst, Jan Griebsch, M.Sc., Prof. Dr. Ernst W. Mayr, Dipl.-Inf. Hanjo Täubig (Technische Universität München). |
The DFG and BMBF projects (BIM, 2+3 years, 1750 TEUR + 1750 TEUR) and BFAM (6 years, 1 full time scientist) fund the establishment of a bioinformatics centre and university programs within the Bioinformatics Initiative Munich, a joint venture of the LMU and TU Munich as well as the GSF research centre and the Max-Planck-Institute for Biochemistry. Overall, BIM funds four junior research groups consisting of one associate professor and two scientists each. The groups are associated with the respective faculties for computer science and biology at the two universities. In the BFAM project a goal is to establish a graduate program for graduates of either computer science/mathematics or biology/chemistry. Here graduates will study via an individual study plan to receive an additional bachelor degree in bioinformatics within three semester terms. At the research unit for practical computer science and bioinformatics, a junior research group (Associate professor Heun (Stiftungsprofessur) and two scientists) is funded via BIM for a five year period (2003-2008).
Related publications: [40, 41, 42, 43]
Members of the Chair have been organisers of the following conferences: German Conference on Bioinformatics 2003, GCB'03, October 12-14, 2003, Neuherberg-Garching. [70]
International Conference on Intelligent Systems in Molecular Biology 1999, ISMB'99, Heidelberg, August 8-12, 1999. [61]
The unit participates in the organisation and program committees of the major national and international bioinformatics conferences (ISMB International Conference on Intelligent Systems in Molecular Biology; RECOMB International Conference on Computational Biology; GCB German Conference on Computational Biology; ECCB European Conference on Computational Biology) and presents papers at these highly competitive conferences. The unit performs several DFG, BMBF, and industry funded projects and co-operates with several other universities (Bonn, Cologne, Erlangen, Frankfurt, Hamburg, Mainz, Münster, Rummelsberg, Saarbrücken), research institutes (caesar, Bonn) and research centres (Fraunhofer society, institute SCAI, Sankt Augustin; DKFZ, Heidelberg; GBF, Braunschweig; and GSF, Neuherberg), Max-Planck-Institutes (MPI Informatics, Saarbrücken; MPI Biochemistry, Martinsried), large companies (Aventis, Frankfurt; Boehringer Ingelheim; Bayer Pharma, Wuppertal), medium (GPC AG, Martinsried; Medigene AG, Martinsried; Qiagen GmbH, Hilden; Bruker Daltronic, Bremen) as well as small start-up companies (BioTeZ GmbH, Berlin-Buch; Invitek, GmbH, Berlin; Nascacell AG, Tutzing, BioSolveIT GmbH, Sankt Augustin). Results of research projects are utilised and commercialized according to the BMBF rules via a bioinformatics start-up company BioSolveIT GmbH. Professor Zimmer co-founded BioSolveIT GmbH, Sankt Augustin, as a spin-off of the GMD/FhG in 2001. Currently, BioSolveIT has created 15 jobs, mostly for bioinformatics researchers at the postdoc level. In 2002, the unit has raised funds of about 1200 TEUR for employment and further qualification of ten scientists and ten students and for equipment (120 TEUR for a server/cluster computer). In addition, the Bioinformatics Initiative Munich is funded with about 3500 TEUR by the DFG.
Publications: [1, 2, 3, 5, 6, 7, 8, 9, 10, 11, 12, 15, 16, 17, 18, 19, 20, 21, 22, 25, 26, 28, 32, 33, 34, 35, 37, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 55, 57, 58, 59, 60, 61, 62, 63, 64, 67, 69, 70, 71, 73, 75, 76, 77, 78, 79, 80, 81, 82, 83, 85, 86, 88, 89, 91, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107]
Poster presentations: [6, 27, 36, 56, 68]
Dissertation Theses: [31, 38, 65, 84, 87, 92]
Bachelor and Master Theses: [13, 14, 23, 24, 26, 29, 30, 54, 66, 72, 74, 90]
University lecture notes: [41, 43]
Various workshops ranging from Semi- and nonparametric Modelling to Statistics in Genetics have been organized.
A large number of researchers have visited the department, various seminars have been given and research projects initiated.
| Sonderforschungsbereich 386 "Statistical Analysis of Discrete Structures" |
€ | 3.765.440,00 |
| Statistical Consulting Unit (1998-2002) | € | 546.808,51 |
| DFG-"Emmy-Noether" Fellowship 2000 - 2001 | € | 70.000,00 |
| DFG-"Emmy-Noether" Research Group (2002-2003) | € | 380.000,00 |
| VW-Project (total volume DM 1.77 Mio.) in Munich since 2003 | € | 597.515,00 |
Methodological research at the department is usually motivated and stimulated through demand and challenge in diverse fields of applications and empirical research. Vice versa, applied research in life, economic and social sciences is often based on our own development of adequate statistical methods or on related work of colleagues in our scientific community.
This interplay between methods and applications is reflected in the structure of research activities, and the following topics are of prime applied or methodological interest.
Statistical Modelling
Computational Statistics
Econometrics
Biostatistics
Statistics in business, economics and social
sciences
Methodological Foundations of Statistics
These research clusters enhance the formation of groups of scientists cooperating within the department. Additionally, this structure stimulates joint scientific work and research seminars, and it aims at transcending traditional organizational borders.
The following sections provide an overview on research activities in these clusters. The clusters are partially overlapping, so some research results appear in more than one topic.
Classical statistical models are generally useful in situations where data are approximately Gaussian and can be explained by some linear structure. Although easy to interpret and theoretically well understood, underlying assumptions are often too restrictive in situations where data are clearly non-Gaussian or have nonlinear structure. Driven by the demands in biological, economic and social sciences, and grown around generalized linear models, statistical modelling emerges as a broad and flexible extension of model-based statistical inference in more complex data situations, in particular with discrete and correlated data. Inference is mostly likelihood-based, including modern Bayesian approaches.
We roughly distinguish three major, overlapping subclusters: Semi- and nonparametric Regression (Likelihood-based Semiparametric Regression and Bayesian Semiparametric Regression), Deficient Data (Missing data and Measurement error models), Time-dependent and Spatial Data (Time-dependent and Spatial data).
Semiparametrically structured regression models are defined as a class of models for which the predictors may contain parametric parts, additive parts with an unspecified functional form of covariates as well as interactions between variables which are described as varying coefficients. The approaches are extremely flexible in capturing the way in which the predictor influences the dependent variable. Research focuses on approaches which are embedded into the framework of semiparametric generalized models, allowing for response variables which are given as count data or binary variables, or metrically scaled variables of various distributional form. Development of methods includes localizing approaches as well as penalized maximum likelihood methods.
Jochen Einbeck, Göran Kauermann, Helmut Küchenhoff, Torsten Scholz, Gerhard Tutz
Dreesman and Tutz (2001), Edlich, Kauermann and Tutz (1998), Einbeck (2001), Einbeck (2003), Einbeck, de André and Singer (2002), Einbeck and Kauermann (2003), Galindo, Kauermann, Liang and Carroll (2000), Kauermann (2000), Kauermann (2002), Kauermann and Berger (2003), Kauermann and Carroll (2000), Kauermann and Opsomer (2001), Kauermann and Tutz (1999), Kauermann and Tutz (2000), Kauermann and Küchenhoff (2003), Kauermann and Tutz (2003), Simonoff and Tutz (1999), Tutz (2000), Tutz (2002a), Tutz (2002b), Tutz (2003a), Tutz (2003b), Tutz, Edlich and Bäumer (1999), Tutz and Kauermann (2003), Tutz and Scholz (2003a), Tutz and Scholz (2003b).
SFB 386, Projekt C4 Local adaptive statistical methods
Bayesian approaches for non- and semiparametric regression models have recently gained much interest. They offer some advantages, in particular: choice of smoothing or tuning parameters are an integral part of the model, and extensions to more complex situations, such as longitudinal or spatial data, are conceptually easy. We distinguish between smoothness prior approaches as a stochastic generalization of penalized likelihood methods and adaptive basis function approaches. We have worked in both directions, but our current focus is on smoothness prior approaches. Inference is usually fully Bayesian using Markov chain Monte Carlo (MCMC) simulation. As an alternative, we developed empirical Bayes inference based on mixed model technology. To make the methods accessible for nonspecialists and to facilitate cooperation with applied researchers, we developed public-domain software, in particular BayesX, see Computational Statistics.
Clemes Biller, Andreas Brezger, Ludwig Fahrmeir, Alexander Jerak, Thomas Kneib, Stefan Lang.
Biller (2000a, b); Fahrmeir (2001); Brezger (2003); Brezger and Lang (2003); Fahrmeir (1999); Fahrmeir and Lang (2001a, b); Fahrmeir and Tutz (2001), Fahrmeir and Mayer (2001), Jerak and Lang (2003); Klinger et al (2000); Klinger (2001); Lang and Brezger (2003); Lang, Adebayo and Fahrmeir (2002); Lang, Fronk and Fahrmeir (2002); Lang (2003).
Our academic collaborators are Michel Denuit (Louvain), Stephan Klasen (LMU, VWL, and University of Göttingen), Dietmar Harhoff (LMU, BWL), Winfried. Steiner (University of Regensburg, Marketing), Kurt Ulm (TUM, IMSE).
Partners in the banking and insurance sector and in other institutions are Allianz Group (Munich), HypoVereinsbank (Munich), SCHUFA (Bonn).
SFB 386, Project C1 “Semi- and nonparametric methods”, since 1995, Humboldt Foundation (Murray Smith).
In practical statistics investigators often are confronted with the problem of incomplete data sets. Therefore, statistics as a field of research has to develop empirical-analytical tools to deal with this problem. In the centre of research we consider the estimation and the prediction in models of regression type ander the complication of incomplete data. This concerns missing data problems for longitudinal- and cluster data as well as linear regression models with incomplete discrete and continuous covariates. The methods partially were implemented in the software tool MAREG & WinMAREG. Further research topics are semiparametric models with missing data, selection models with flexible modelling of the drop-out rate and generalized linear models with random effects and missing MNAR response. In non- and semiparametric regression model methods known from linear regression and the Nearest Neighbor Imputation were investigated. Marginal regression models, conditional models and random effects models are a possible adaption of Generalized Linear Models to dependent response and were also handled.
C. Kastner, S. Storck, T. Nittner, S. Scheid, N. Lieske, A. Dörfler, A. O. Adejumo (DAAD)
Fahrmeir, Gieger and Heumann (1999), Fieger (2001), Heumann (1998), Kastner (2001), Nittner (2003), Rao, Srivastava and Toutenburg. (1998), Rao and Toutenburg (1998), Rao and Toutenburg (1999), Rao and Toutenburg (2001), Shalabh and Toutenburg (2003), Toutenburg (2002), Toutenburg (2002), Toutenburg, Fieger and Heumann (2000), Toutenburg, Fieger and Schaffrin (2002), Toutenburg, Fieger and Srivastava (1999) Toutenburg, Heumann, Nittner and Scheid (2002), Toutenburg and Shalabh (2000), Toutenburg and Shalabh (2001), Toutenburg and Shalabh (2002), Toutenburg and Shalabh (2003), Toutenburg and Srivastava (1998), Toutenburg and Trenkler (1998), Toutenburg and Nittner (2002), Toutenburg and Srivastava (2001), Toutenburg and Srivastava (2002), Toutenburg, Srivastava, Schaffrin and Heumann (2003), Ziegler, Blettner, Kastner and Chang-Claude (1998), Ziegler, Hebebrand, Kastner and Müller-Myhsok (1999), Ziegler, Kastner and Blettner (1998).
Research cooperation depends on projects currently having problems with missing data. In the past these were, for example, the projects B6, B7, C1 and C4 of SFB 386.
SFB 386 (C3 "Missing Data", 1995-2003), DAAD
A widespread problem in applying regression analysis is the presence of measurement error. Often the variables of interest cannot be observed directly or measured correctly, and one has to be satisfied with surrogates (often also named indicators or proxies). If one ignores the difference between the ideal variables in the model and their observable counterparts, and just plugs in the surrogates instead of the variables (`naive estimation'), then all the estimators must be suspected to be severely biased. So the development of adjusted estimators is indispensable to avoid deceptive conclusions. These methods have received increasing attention especially in epidemiology and econometrics.
One focus of our research is survival analysis, where we, in particular, derived an exact corrected score function for the Cox model and a general unifying approach to deal with measurement error in parametric accelerated failure time models. Further research was concerned with the comparison of different approaches with regard to efficiency, the behavior of structural estimators under misspecification, and the search for robust (outlier resistant and Ridge type) estimators. We also work on nonstandard measurement error problems in linear and nonlinear models, like rounding, heaping, complex error models, the superposition of Berkson and classical error, and deliberately contaminated data to guarantee anonymity.
We have applied the methods in nutritional epidemiology (influence of nutrition habits on cardiovascular disease), several radiation studies, micro-econometrics and sociology (unemployment duration data from the German Socio-Economic Panel) and dental medicine
A. Kukush, I. Langner, D. Rummel, M. Thamerus, R. Schwarz, G. Schuster, R. Wolf, T. Nittner.
Augustin (2000); Augustin (2002c); Augustin (2002d); Augustin (2003a); Augustin, Döring and Rummel (2003); Augustin and Wolff (2003); Augustin and Schwarz (2002); Bender, Augustin and Blettner (2003); Cheng and Schneeweiss (1998); Cheng and Schneeweiß (2002); Cheng, Schneeweiß and Thamerus (2000); Heid, Gerken, Wellmann, Küchenhoff, Kreienbrock and Wichmann (2002); Heid, Schaffrath, Kreienbrock, Küchenhoff and Wichmann (2003); Küchenhoff, Bender, Langner and Lenz-Tönjes (2003); Kukush and Maschke (2003); Kukush, Schneeweiß and Wolf (2003); Kukush, Schneeweiß and Wolf (2001a); Kukush, Schneeweiss and Wolf, R. (2001b); Kukush, Schneeweiss and Wolf (2002); Schneeweiß (2003); Schneeweiß and Cheng (2003); Schneeweiß, Cheng and Wolf (2002); Schneeweiß and Nittner (2001); Shklyar and Schneeweiß (2002); Thamerus (1998); Thamerus (2003); Wolff and Augustin (2003);
Ahmad Amar, John Komlos, Marco Sander (Institute of Economic History, University of Munich), A. Döring and H.-E. Wichmann, I. Heid, GSF–National Research Center for Environment and Health, Neuherberg), M. Blettner and R. Bender, School of Public Health, University of Bielefeld), Institut für Angewandte Wirtschaftsforschung, Tübingen, A. Kukush (Kiev), E. Lesaffre (Kath. University of Leuven), Chi-Lun Cheng (Academia Sinica, Taipei), Raymond Carroll (Texas A&M), Ori Davidov (Haifa, Israel).
SFB 386, Project C2: "Mess- and Klassifikationsfehler mit diskreten Daten". Health Effects Institute (USA).
In life and social sciences as well as in business and industry, the availability of data that carry temporal or spatial information is nearly exploding and creates an important and challenging topic of current international research where the Department contributes. Our methodological research is mainly motivated and driven by biostatistical and economic applications, in cooperation with partners from various fields. The focus is on models and methods for longitudinal data, in particular with discrete responses, for survival and event history data, and for spatial or spatio-temporal data. The complexity of realistic models for temporal and spatial data necessitates computer-intensive data analytic methods, thus strengthening the link to Computational Statistics.
Approaches with an emphasis on economic times series and longitudinal data, in particular for data from financial markets, are contained in Econometrics.
Clemens Biller, Ludwig Fahrmeir, Leonhard Held, Thomas Kneib, Stefan Lang, Andrea Hennerfeind, Leyre Osuna, Günter Raßer, Volker Schmid, Gerhard Tutz.
Biller (2000c); Biller (2001); Dreesman and Tutz (2001); Fahrmeir, Gieger and Heumann (1999); Fahrmeir and Künstler (1999); Fahrmeir and Knorr-Held (2000); Fahrmeir and Hennerfeind (2003); Fahrmeir, Gössl and Hennerfeind (2002, 2003); Fahrmeir, Kneib and Lang (2003); Gehrmann et al. (2003); Gieger (1999); Hennerfeind, Brezger and Fahrmeir (2003); Kauermann and Tutz (2001); Knorr-Held (2000); Knorr-Held and Besag (1998); Knorr-Held and Raßer (2000); Lang and Brezger (2003); Schmid and Heald (2003); Tutz (2002); Tutz (2003b); Tutz and Binder (2002); Yassouridis et al. (1999).
Dorothee Auer (MPI für Psychiatrie), Nicole Augustin (University of Glasgow), German Cancer Research Center (DKFZ, Heidelberg), Stephan Klasen (VWL, LMU und Universität Göttingen), Monica Musio (FDM, Freiburg), Albrecht Neiss (IMSE, TUM and Sylvia Lawry Center for Multiple Research); Michael Smith (University of Sydney, Australia); Kurt Ulm (IMSE, TUM).
SFB 386, Project A5 „Spatial Statistics“ since 1998; Humboldt Foundation (Michael Smith); Graduiertenkolleg Angewandte Algorithmische Mathematik (TUM).
Computational statistics is a statistical science at the interplay between computer science and data analysis. The topic includes various state-of-the-art methods for statistical inference such as resampling methods (e.g. bootstrap and jackknife), Monte Carlo and Markov chain Monte Carlo (MCMC) methods, non- and semiparametric regression as well as methods for prognosis, classification and data mining.
MCMC methodology provides enormous scope for realistic statistical modelling. Research at the department has focused on designing efficient MCMC algorithms for latent parameter in complex hierarchical models. In particular, methods for estimating latent Gaussian Markov random fields (GMRFs) and Bayesian P-Spline models have been developed, with strong emphasis on so-called block updating algorithms. Such algorithms have considerably improved convergence and mixing properties. Furthermore, methods based on auxilliary variables have been investigated, which allow for block-updating via Gibbs sampling in binary and multicategorical regression problems, in contrast to Metropolis-Hastings steps based on multivariate Taylor expansions. Finally, specific implementations of Bayesian partition models via reversible jump MCMC have been developed. The algorithms are basic building blocks for fully Bayesian inference in complex semiparametric models, see 3.1 “Statistical Modelling”.
Andreas Brezger, Ludwig Fahrmeir, Leonhard Held, Andrea Hennerfeind, Alexander Jerak, Stefan Lang, Leyre Osuna, Günter Raßer, Volker Schmid
Brezger and Lang (2003); Fahrmeir and Lang (2001a); Fahrmeir and Lang (2001b); Fahrmeir and Hennerfeind (2003); Fahrmeir and Osuna (2003); Guidici, Knorr-Held and Rasser (2000); Held (2004); Hennerfeind, Brezger and Fahrmeir (2003); Holmes and Held (2003); Jerak and Lang (2002); Jerak and Lang (2003); Knorr-Held (1999); Knorr-Held and Raßer (2000); Knorr-Held and Rue (2002); Lang and Brezger (2003); Lang, Fronk and Fahrmeir (2002); Lang, Adebayo and Fahrmeir (2002); Lang, Adebayo, Fahrmeir and Steiner (2003); Schmid and Held (2003).
Chris Holmes (Imperial College London), Havard Rue (Norwegian University of Science and Technology, Trondheim).
SFB, project A5 “Spatial Statistics” and C1 “Non- and semiparametric Methods”, since 1995.
Scientific statistical computing requires reliable statistical software. We test standard statistical software packages and point out possible errors so that these errors can be eliminated in future versions of the software. Great efforts have been taken to provide public domain software for the new statistical methodology developed at the department. The following statistical packages have been developed (among others):
Bamp (www.stat.uni-muenchen.de/~schmidt/bamp/): A program for Bayesian Age-Period-Cohort-Modelling and Prediction.
BayesX (www.stat.uni-muenchen.de/~lang/bayesx/bayesx.html): Software for Bayesian Inference. BayesX is able to estimate very complex semiparametric regression models with sutructured additive predictors in a Bayesian framework.
BDCD (www.stat.uni-muenchen.de/~rasser/bdcd/): A program for Bayesian Detection of Clusters and Discontinuities in Disease Maps. This command-line based program allows for the estimation of unknown relative risk parameters in a typical disease mapping setting.
BVCM (www.stat.uni-muenchen.de/sfb386/software/bvcm.html): A program for Bayesian Varying-Coefficient Models. The software estimates varying coefficient models in a Bayesian framework.
ELV (www.stat.uni-muenchen.de/~knuesel): Computation of elementary statistical distributions.
GENcon/GENbin: Programs for solving optimization problems characterized by real respectively binary values. Both software tools base on the concept of genetic algorithms and run on MATLAB6.
GraphFitI ( www.stat.uni-muenchen.de/~blauth/GraphFitI/graphFitI.html): A program for model selection in graphical chain models (Graphical models Fitting Interactions). It is designed to fit a graphical model to a multivariate data set; it applies a data-driven selection strategy introduced by Cox and Wermuth.
Mareg/WinMareg ( www.stat.uni-muenchen.de/~andreas/mareg/winmareg.html): A Program for computation of semiparametric (GEE1, GEE2) and likelihood based marginal models for longitudinal and clustered categorical data with treatment of missing values.
S-Plus Code for multicategorical penalized spline regression. Based on the P-Splines approach the software allows for nonparametric extensions of common models for nominal and ordinal responses.
For further software projects see the section about Statistical Genetics and Bioinformatics.
Christiane Belitz, Angelika Blauth, Clemens Biller, Andreas Brezger, Angelika Caputo, Vanessa Didelez, Eva-Maria Fronk, Leonhard Held, Andrea Hennerfeind, Manuela Hummel, Alexander Jerak, Stefanie Kessner, Thomas Kneib, Leo Knüsel, Petra Kragler, Rüdiger Krause, Stefan Lang, Leyre Osuna, Iris Pigeot, Günter Raßer, Evi Rainer, Volker Schmid, Torsten Scholz, Kurt Watzka.
Biller (2000a, 2000b, 2000c); Blauth (2002); Blauth, Pigeot and Bry (2000); Brezger, Kneib and Lang (2003a, 2003b); Fahrmeir, Gieger and Heumann (1999); Fahrmeir, Kneib and Lang (2003); Fieger, Kastner and Heumann (1998); Heumann (1998, 2000, 2003); Heumann, Fieger and Kastner (1998); Kastner, Fieger and Heumann (1999); Knorr-Held and Rainer (2001); Knorr-Held and Raßer (2000); Knüsel (2001, 2002, 2003a, 20003b); Krause and Tutz (2003a); Toutenburg, Fieger and Heumann (1998, 1999); Toutenburg et al. (2002, 2003); Trevisani et al. (1999); Tutz and Scholz (2003a).
Multivariate statistical analysis is concerned with data that consist of sets of measurements on a number of individuals or objects. The basis is the analysis of dependence between variables, between sets of variables and between variables and sets of variables. The investigation of the structure of variables is used in prognosis and classification and for the detection of similarity of objects. Methods are often based on computer intensive methods as boosting or bootstrapping, genetic algorithms and tree-based methods. In high dimensional statistical analysis problems of dimension reduction prevail. Strongly related to multivariate methods is statistical data mining which is the process of selecting, exploring, modifying and modelling large sets of data by statistical methods to uncover previous unknown patterns. Bump hunting has been explored as a new tool of statistical data mining for analyzing risks in finance and survival analyis.
Anne-Laure Boulesteix, Jochen Einbeck, Rüdiger Krause, Ludwig Fahrmeir, Klaus Hechenbichler
Becker and Fahrmeir (2000, 2001), Boulesteix and Tutz (2003), Einbeck (2003), Einbeck and Kauermann (2003), Einbeck, Tutz and Evers (2003), Friedl and Kauermann (1999), Galindo, Kauermann, Liang and Carroll (2000), Kauermann and Opsomer (2001), Krause and Tutz (2003a,b), Spieß and Tutz (2003), Tutz and Binder (2003), Tutz and Evers (2003), Tutz and Hechenbichler (2003).
Deutsches Institut für Wirtschaftsforschung (Berlin).
SFB 386, Projekt C4 Local adaptive statistical methods
Graphical models are a useful tool for modelling complex high-dimensional association. The group has developed theoretical and computational tools in order to apply graphical models in a wide area of applications. Furthermore, extensive research has been done in applying resampling methods in the area of bioaquivalence and non-inferiority trials.
Angelika Blauth, Angelika Caputo, Vanessa Didelez, Eva-Maria Fronk, Astrid Heinicke, Iris Pigeot-Kübler, Kurt Watzka, Astrid Zierer
Blauth, Pigeot and Bry (2000); Caputo, Foraita, Klasen and Pigeot (2003a); Caputo, Heinicke and Pigeot (1999); Didelez (2002, 2003); Didelez and Pigeot (1998); Didelez, Pigeot, Dean and Wister (2002); Fried and Didelez (2003); Lang, Fronk and Fahrmeir (2001); Pigeot, Heinicke, Caputo and Brüderl (2000); Shao, Kübler and Pigeot (2000).
SFB 386, Project B5
Econometrics combines economic theory with techniques from statistics and mathematics to model economic and financial systems. Econometric models help to better understand the economic processes we observe and, thus, to improve the design of economic systems. Specifically, econometric models play an important role in testing economic theories, in predicting future economic developments and in supporting economic policy making. The department’s research activities focus particularly on econometric methods, financial econometrics, time series analysis, empirical macroeconomics, and empirical industrial economics.
The department’s research in the area of econometric methods covers a range of issues pertaining to estimation, testing and economic decision making. Among others, it includes the analysis of regression models in the presence of heavy-tailed disturbances, testing for the presence of structural breaks, and the detection of outliers.
Erich-Otto Maschke
Schneeweiß, H., Maschke, E.-O., Pfannes, M. (2001); Klebanov et al. (2000); Mittnik, S., Rachev, S.T. (2000); Mittnik, S., Paulauskas, V., Rachev, S.T. (2001); Mittnik, S., Rachev, S.T., Samorodnitsky, G. (2001); Klein, I., Mittnik, S. (2002); Mittnik, S., Rachev, S.T., Samorodnitsky, G. (2000); Mittnik, S., Rachev, S.T. (2001).
Compared to other economic data, financial data often behave very differently. Specifically, they can be characterized by heavy tails (i.e., extreme price movements are more likely than the normal distribution implies) and, although practically uncorrelated over time, can exhibit strong temporal depencencies in higher moments, such as found in volatility. These properties prevent the use of most “standard” techniques developed in statistics and econometrics for a wide range of questions in empirical finance. The department’s research in this area is concerned with both methodological developments as well as practical applications, such as the prediction of financial risk, the examination of market efficiency, portfolio optimization, and modeling energy markets.
Toker Doganoglu, Markus Haas, Christoph Hartz, Christian Pigorsch, Corinna Prange
Claessen, H., Mittnik, S. (2002); Doganoglu, T. Mittnik, S. (2003); Haas, M., Mittnik, S., Paolella, M.S. (2002); Haas, M., Mittnik, S., Paolella, M.S. (2003); Hartz, C., Mittnik, S., Paolella, M. (2003a); Hartz, C., Mittnik, S., Paolella, M. (2003b); Küster, K., Mittnik, S., Paolella, M. (2003); Mittnik, S., Paolella, M.S. (2000); Mittnik, S., Paolella, M.S., Rachev, S.T. (2000); Mittnik, S., Rachev, S.T. (2000); Mittnik, S., Rieken, S. (2000); Mittnik, S., Rieken, S. (2000); Mittnik, S, Rachev, S.T., Samorodnitsky, G. (2001); Mittnik, S., Paolella, M.S., Rachev, S.T. (2002); Mittnik, S, Rachev, S., Schwartz, E. (2002); Mittnik, S., Paolella, M.S. (2003); Hübler, O., Mittnik, S., Schmid, W. (2002); Mittnik, S. (2001); Mittnik, S., Rachev, S.T. (2001); Mittnik, S., Rachev, S.T. (2002); Pigorsch, C. (2003); Prange, C. (2003a); Prange, C. (2003b)
Svetlozar Rachev (University of Karlsruhe and University of California – Santa Barbara), Genady Samorodnitsky (Cornell University), Marc Paolella (University of Zurich), Francis X. Diebold (University of Pennsylvania), Eduardo Schwartz (University of California – Los Angeles).
Center for Financial Studies, Frankfurt; Institut für Quantitative Finanzanalyse GmbH, Kiel; HSH Nordbank, Hamburg and Kiel.
The research in time series analysis addresses both, theoretical and applied issues. In addition to economic and financial applications, there are contributions in the area of medical psychology.
Markus Haas, Christoph Hartz
Haas, M., Mittnik, S., Paolella, M.S. (2002); Haas, M., Mittnik, S., Paolella, M.S. (2003); Hartz, C., Mittnik, S., Paolella, M. (2003a); Mittnik, S., Paolella, M.S., Rachev, S.T. (2003); Schubert Ch. et al. (2001); Mittnik, S., Rachev, S.T., Samorodnitsky, G. (2000); Schubert, Ch. et al. (2003).
Christian Schubert (University Hospital, Innsbruck), Marc Paolella (University of Zurich), Svetlozar Rachev (University of Karlsruhe and University of California – Santa Barbara), Peter Zadrozny (Bureau of Labor Statistics, Washington, D.C.).
The research in empirical macroeconomics investigates the interaction of the financial sector and the real sector of an economy. A second field of investigation concerns questions in public economics. Here, we analyze the effect public spending, specifically public consumption versus public investment has an long term economic growth.
Chiarella, C. et al. (2002); Mittnik, S., Neumann, T. (2001); Mittnik, S., Neumann, T. (2003).
Thorsten Neumann (Deka Bank, Frankfurt), Willi Semmler (University of Bielefeld and New School, New York), Peter Zadrozny (Bureau of Labor Statistics, Washington, D.C.).
ifo Institute for Economic Research, Munich.
In the last two decades, Industrial Economics have been going through an empirical revolution. Researchers have applied econometric techniques beyond hypothesis testing, and as a toolbox for measurement of important industry determinants. By means of structural models, important unobservables such as demand elasticities and marginal costs can be measured. Furthermore, newly developed demand estimation techniques allow measurement of impact of policies on welfare and costs, thus econometric analysis has become a necessary step in antitrust and merger analysis. The group employs such techniques to analyze network industries, most notably telecommunications. The research is concentrated on estimating demand in a competitive telecommunications industry, using these to measure welfare impact of liberalization, and on measuring competitiveness of industries across Europe. Other research areas are internet auctions, network interconnection, research and development with network effects, and open source software development.
Daniel Cerquera, Sascha Frohwerk, Lukasz Grzybowski, Kerstin Lange, Martin Reichuber, Katharina Sailer
Doganoglu, T. (2002); Doganoglu, T. (2003); Sailer, K. (2001); Doganoglu, T., Sailer, K. (2003); Klodt, H. (2003); Sailer, K. (2001), Doganoglu, T. (2003a); Doganoglu, T.(2003b); Doganoglu, T., Wright, J. (2003); Doganoglu, T., Grzybowski, L. (2003); Grzybowski, L. (2003); Lange, K. (2002); Lange, K. (2003); Frohwerk, S. (2003).
Institute for World Economics, Kiel
Volkswagen Foundation.
Biostatistics creates and applies methods for quantitative research in the health sciences. Common applications include clinical medicine, epidemiological studies, genetics, environmental health, ecology, forestry, and general biology. At the Department of Statistics at Munich University particular emphasis is placed on research related to epidemiology, genetics, and neuro science.
Research is often done in close collaboration with various national and international partners.
The group develops methodology for spatial and spatio-temporal, longitudinal, and survival data on chronic and infectious diseases. Recent work has focussed on the spatial and temporal analysis of cancer incidence and mortality. Furthermore, problems in infectious disease epidemiology of animals and humans are being considered. Papers on these issues in journals like "Biometrics", "Biostatistics", "Statistics in Medicine" and "Applied Statistics" have led to international recognition of the group.
Leonhard Held, Michael Höhle, Mathias Hofmann, Argwings Otieno Ranyimbo, Volker Schmid.
Carvalho and Knorr-Held (2003); Crook, Knorr-Held and Hemingway (2003); Diggle, Knorr-Held, Rowlingson, Su, Hawtin and Bryant (2003); Guidici, Knorr-Held and Raßer (2000); Hansell, Knorr-Held and Schmid (2002); Hansell (2003), Held, Natario, Fenton, Rue and Becker (2003);Höhle (2002); Höhle, Jørgensen and O'Neill (2003); Knorr-Held (2000, 2003); Knorr-Held and Becker (2000); Knorr-Held and Besag (1998); Knorr-Held and Best (2001); Knorr-Held and Rasser (2000); Knorr-Held, Raßer and Becker (2002); Knorr-Held and Rainer (2001); Knorr-Held and Richardson (2003); Natario and Knorr-Held (2003); Schmid and Held (2003); Staubach, Schmid, Knorr-Held and Ziller (2002).
Our partners at research institutes and universities are Nikolaus Becker (German Cancer Research Center (DKFZ), Heidelberg), Robert-Koch-Institut (RKI), Berlin, Christoph Staubach (Department of Epidemiology at the Federal Research), Centre for Virus Diseases of Animals, Wusterhausen), Peter Diggle (Medical Statistics Group at Lancaster University), Nicky Best, Angela Crook, Anna Hansell, and Sylvia Richardson (Department of Epidemiology and Public Health at Imperial College London), Philip O'Neill (Nottingham University, Statistics Division), Marilia Sa Carvalho (National School of Public Health/FIOCRUZ, Rio de Janerio, Brasilien), Isabel Natario (University of Lisbon)
SFB386, Project A5 “Spatial Statistics” (since 1995), B9 “Statistical Methods for Surveillance of Infectuous Diseases” (since 2003), DAAD (Argwings Otieno Ranyimbo).
Anne-Laure Boulesteix, Gangolf Jobb, Rainer Opgen-Rhein, Juliane Schäfer, Korbinian Strimmer, Gerhard Tutz, Sofia Wichert.
Paradis et al. (2003); Boulesteix et al. (2003); Wichert et al. (2003); Strimmer et al. (2003); Strimmer (2003); Schmidt et al. (2002); Strimmer and Rambaut (2002); Strimmer and Pybus (2001); Drummond and Strimmer (2001); Strimmer et al. (2001); Salemi et al. (2001); Strimmer and Moulton (2000).
The group actively develops a number of software package for the analysis of genetic data, that are freely available for download:
APE: R-package for statistical analyses in phylogenetics and evolution (http://cran.r-project.org/src/contrib/PACKAGES.html#ape).
GeneTS: R package for the analysis of microarray time series data (http://www.stat.uni-muenchen.de/~strimmer/genets/).
TREE-PUZZLE: Program to infer phylogenetic trees from DNA sequences using a maximum-likelihood approach (http://www.tree-puzzle.de).
PAL: Java library for algorithms and methods used in molecular evolution, statistical genetics, and sequence analysis (http://www.cebl.auckland.ac.nz/pal-project/).
TREEFINDER: Java/C program for molecular phylogenetics (http://www.treefinder.de).
The main interest of this group are statistical aspects of bioinformatics and computational biology. Current research focuses on the development of methods for analysing gene expression data and on probabilistic models for DNA sequence analysis.
Local collaboration with experimental and theoretical groups, e.g. at the Center for Nanoscience CeNS/LMU, via the "Munich Systems Biology Forum" (http://www.sysbio-muenchen.de). International cooperation with the Bioinformatics Center (Birc) at the University of Aarhus, and the Biomedical Center of the University of Uppsala.
DFG (Emmy Noether), SFB 386
An important area of substantive research in neuroscience is functional neuroanatomy of the human brain. Brain mapping aims at detecting areas of functional activities, for example the visual cortex, based on functional magnetic resonance imaging (fMRI) data. Compared to conventional methods of “statistical parametric mapping” (SPM), we develop and apply modern Bayesian techniques for modelling and analysing these massive spatio-temporal fMRI data. More recently, functional connectivity, that is detection and tracking of fiber bundles connecting functional areas, based on diffusion tensor imaging (DTI) data, has gained much interest. Because of the structure of fMRI and DTI data, statistical methods in human brain research are strongly related to the research cluster.
Dorothee Auer, Ludwig Fahrmeir, Christoff Gössl, Susanne Heim
Auer et al. (2001); Fahrmeir and Gössl (2002); Gössl, Auer and Fahrmeir (2000); Gössl and Fahrmeir (2001); Gössl, Auer and Fahrmeir (2001); Gössl, Fahrmeir and Auer (2001); Fahrmeir, Gössl and Hennerfeind (2002); Gössl (2002); Gössl et al. (2002); Smith et al. (2003); Fahrmeir, Gössl and Hennerfeind (2003).
Klaus Hahn, Gerhard Winkler (GSF), Mike Smith (University of Sydney, Australia)
SFB 386, Project A5 “Spatial Statistics”, since 1998.
This research cluster comprises primarily applied research and empirical analyses, with an emphasis on tackling and solving substantive problems in business, economics and social science. Topics cover a broad field of applications, ranging from marketing research, industrial economics, risk management for banks and insurance companies, socio-economic development, labour market analyses, official statistics and demography to empirical sociology and psychology. Financial econometrics is subsumed in the cluster Econometrics.
Research is often carried out in joint work with partners from universities and research institutions, or in cooperation with partners in business and industry. Research questions also emerge from consulting cases handled by the Statistical Consulting Unit. The focus is on solving substantive research questions with state-of-the-art methods developed within the department or in related work.
Empirical research and practical problems in business and economics confront scientists and practitioner with increasingly large and complex data sets, requiring modern statistical tools for adequate data analysis. Consulting cases and cooperation’s with partners inside and outside of the university initiate much of our methodological research, and – as an ultimate goal – result in interdisciplinary work and publications. Areas of current major interest are: risk analyses in the credit and actuarial sector, marketing research, labour market analyses, and public health as well as socio- economic problems in developing countries.
Samson Adebayo, Ludwig Fahrmeir, Alexander Jerak, Ngianga-Bakwin Kandala, Stefan Lang, Helmut Küchenhoff, Gerhard Tutz, Renata Eva Zambrzycka.
Adebayo (2003, 2003a, b); Adebayo and Fahrmeir (2003), Bender, Fahrmeir and Lang (2002); Bender, Kohlmann and Lang (2003); Berger, Fahrmeir and Klasen (2002); Denuit and Lang (2003); Fahrmeir and Lang (2003a, b); Fahrmeir, Lang and Spies (2003); Fahrmeir, Lang, Wolff and Bender (2003); Frank (2000); Fahrmeir, Henking and Hülst (2003); Jerak and Wagner (2003); Kandala, Lang and Klasen (2002); Kandala et al. (2001); Kandala (2002), Kauermann and Tutz (2001), Lang et al. (2002), Lang and Sunder (2003), Payer and Küchenhoff (2003), Tutz (1998), Uehlein and Küchenhoff (2003).
Our academic collaborators are Stephan Klasen (VWL, LMU and University of Göttingen, Dietmar Harhoff (BWL, LMU), Winfried Steiner (University of Regensburg), Michel Denuit (Université Louvain).
Partners in the banking and insurance sector and in other institutions are Stefan Bender (IAB, Nürnberg), Allianz (Munich), Hypo-Vereinsbank (Munich), Deutsche Bank (Frankfurt), Deutsche Telekom (Bonn), Deutsche Bahn (Munich), LV von 1871 (Munich), Oberfinanzdirektion (Hannover), SCHUFA (Bonn), System Analysis Dept. Riso National Laboratory (Roskilde/Denmark).
SFB 386, Projects A5 “Spatial Statistics” and C1 “Non- and semiparametric Methods”, since 1998; DAAD; BMBF (starting in 2004); Klapproth-Stiftung.
There is an intensive participation of our department of statistics in empirical research on fields of social and psychological problems. Bernhard Rüger is a member and consultant of different research groups on this field: DPV (German psychoanalytical association), Sigmund Freud Institut Frankfurt and the group “children of world war II in Europe coming to age”: Within these cooperations new methods result in sampling design, hypotheses testing and cohorts analysing.
Rüger (1998a); Kleindienst, Greil, Rüger and Möller (1999); Leuzinger-Bohleber, Stuhr, Rüger and Beutel (2000); Leuzinger-Bohleber, Stuhr, Beutel, Rüger and Kächele (2001); Leuzinger-Bohleber, Stuhr, Rüger and Beutel (2001a, b); Rüger (2001); Leuzinger-Bohleber, Stuhr, Rüger and Beutel (2002); Rüger (2002a); Augustin and Wolff (2003); Wolff and Augustin (2003).
Thomas Augustin
Marianne Leuzinger-Bohleber (SFI, Frankfurt), Hartmut Radebold (University Kassel), Elmar Brähler (University Leipzig), Gereon Heuft (University Münster), Andreas Kruse (DZFA, Heidelberg), Sozialwissenschaftliches Institut, SIM-ISW (Munich).
In order to derive results of relevance and reliable conclusions, statistics and probability indispensably requires rigorous and steady reflection on its methodological foundations. Two subclusters of vivid research at the department can be distinguished: Foundations of Statistical Inference, and Interval Probability .
“Foundations of Statistical Inference” or “Foundations of Probability” usually have a twofold connotation: One connotation points to the task of elucidating, enhancing and formulating the mathematical background of the methods in question. The other points to what may be called the conceptual background and paradigms prevalent in various interpretations of probability. Foundational questions of the second sort arise at the borderline of Statistics, Probability Theory, Philosophy, and Philosophy of Science. There are several – partially conflicting – paradigms how to learn from a sample about the population, i.e. how to make statistical inference. Currently, research in that field is centered around testing statistical hypotheses, statistical estimation theory and modelling genuinely indeterministic frames. Developments in Philosophy of Science contribute to the understanding of probability and inference.
Thomas Augustin, Kurt Weichselberger
Augustin (1998); Augustin (1999a); Augustin (1999b); Augustin (2002a); Pöhlmann and Augustin (2001); Rüger (1998); Rüger (1999); Rüger (2002); Wassmer and Rüger (1998); Weichselberger und Augustin (1998a)
The research group has been devoted to the development of the theory of interval-probability since more than ten years. This theory constitutes a consistent reaction to the methodological debate of the last decades concerning the appropriate description of uncertainty in those situations where the traditional “classical” theory of probability proves to be not suitable - especially because it is not possible to characterize the events by precise numbers distinguishing the respective component of probability. The theory of interval-probability represents a comprehensive generalization of the classical theory, as determined by Kolmogorov’s axioms, conceived to be interpretation-independent. It constitutes a system of axioms and definitions producing statements of the same fundamental rigour as the classical theory – but with a much wider area of application. Therefore it promises to be advantageous in all discipline employing descriptions of uncertainty, especially in economics – e.g. risk analysis and insurance – , in social sciences, in medicine and other empirical fields. Beyond foundations of the theory of interval-probability, its application to statistical inference, to decision theory and to robust statistics may be emphasized as prominent subjects of engagement for the research group.
Anton Wallner, Bernhard Rüger
Augustin (1998); Augustin (1999a); Augustin (1999b); Augustin (1999d); Augustin (2001); Augustin (2002a); Augustin (2002b); Augustin (2003b); Augustin (2003c); Augustin (2003d); Augustin and Coolen (2003); Augustin and Pöhlmann (2003); Rüger (2002), Utkin and Augustin (2003a); Utkin and Augustin (2003b); Wallner (2002); Wallner (2003a); Wallner (2003b); Weichselberger (2000); Weichselberger (2001); Weichselberger (2002); Weichselberger and Augustin (1998b); Weichselberger and Augustin (2003).
Our academic collaborators are Frank Coolen, Dept. of Mathematical Sciences, University of Durham (England); Gert de Cooman, SYSTeMS Research Group, Ghent University (Belgium); Renato Pelassoni and Paolo Vicig, Dept. of Applied Mathematics, Faculty of Economics, University of Trieste (Italy); Lev Utkin, Depts. of Computer Sciences and Mathematical Modeling in Economics, St. Petersburg Forest Technical State Academy (Russia).
Partners in the banking and insurance sector and in other institutions are Martin Gümbel, izb-software (computing center of the Sparkassen bank), Munich; Igor Kozine, System Analysis Dept. Riso National Laboratory, Roskilde (Denmark); Sigrid Pöhlmann, LV von 1871 (insurance company), Munich.
From July 1999 to June 2000: Anton Wallner as reseach assistant by DFG.
In 2001 the publication of the first volume of the book Elementare Grundbegriffe einer allgemeineren Wahrscheinlichkeitstheorie was supported by DFG.
Workshop Recent Developments in the Theory and Application of Interval-Probability, Munich, May 1st, 2002, partially sponsored by the DFG and from the NATO and London Mathematical society travel grants.
Academia Sinica, Taipei; Bioinformations Center, University of Aarhus; Biomedical Center, University of Uppsala; BWL, University of Munich; Bureau of Labour Statistics, Washington D.C.; Center of Nanoscience, University of Munich; Center of Virus Diseases of Animals, Wusterhausen; Cornell University; Davidov Ori, Haifa/Israel; Denuit Michel, Louvain; Department of Applied Mathematics, University of Trieste/Italy; Department of Computer Sciences and Mathematical Modelling in Economics, St. Petersburg Forest Technical State Academy/Russia;Department of Mathematical Science, University of Durham; Deutsches Institut für Wirtschaftsforschung, Berlin; German Cancer Research Center, Heidelberg; DZFA, Heidelberg; GSF-National Research Center for Environment and Health/Neuherberg; Imperial College, London; IMSE/TU Munich; Institut für Angewandte Wirtschaftsforschung, Tübingen; Institute of Economic History, University of Munich; Institute for World Economics, Kiel; Katholic University of Leuven; Kukush A., Kiev; Marketing, University of Regensburg; Max-Planck-Institute for psychatry, Munich; Medical Statistics Group, University of Lancaster; National School of Public Health, Rio de Janeiro/Brasilia; New School, New York; Norvegian University of Science and Technology, Trondheim; Nottingham University; Robert-Koch-Institute, Berlin; School of Public Health, Siegmund Freud Institute, Frankfurt; Sylvia Lawry Center for Multiple Research, Munich; Systems Research Group, Ghent University/Belgium; Texas A & M.; University of Bielefeld; University of Santa Barbara/California; University of Glasgow; University Hospital, Innsbruck; University of Kassel; University of Karlsruhe; University of Leipzig; University of Los Angeles/California; University of Lisbon; University of Munster; University of Pennsylvania; Universidade de Sao Paulo; University of Sidney/Australia; University of Zurich; VWL/University of Munich and University of Göttingen.
Allianz Group (Munich), Center of Financial Studies (Frankfurt), Deka Bank (Frankfurt), Deutsche Bank (Frankfurt), Deutsche Telekom (Bonn), Deutsche Bahn (Munich), HSH Nordbank (Hamburg/Kiel), HypoVereinsbank (Munich), IAB (Nürnberg), ifo Institute for Economic Research (Munich), Institut für Quantitative Finanzanalyse GmbH (Kiel), izb-Software (Munich), Oberfinanzdirektion (Hannover); SCHUFA (Bonn).
Marit Holden (Oslo); Mark Berliner (Columbus/Ohio); Nicola Best (London); Rainer Dahlhaus (Heidelberg); Arnoldo Frigessi (Oslo); Katja Ickstadt (North Carolina); Hans Künsch (Zürich); Diana Miglioretti (Maryland); Antonio Pievatolo (Mailand); Jörg Polzehl (Berlin); Vladimir Spokoiny (Berlin).
Adrian Bowman (Glasgow); Leo Breiman (Berkeley); Raymond Carroll (Texas A&M); Holger Dette (Bochum); Wolfgang Härdle (Berlin); Robert Kohn (Sydney); Burkhard Seifert (Zürich); Sara van Geer (Leiden); Lue Ping Zhao (Seattle).
Rainer Dahlhaus (Heidelberg); David Edwards (Kopenhagen); Ursula Gather (Dortmund); Zhi Geng (Peking); Paolo Giudici (Pavia); Niels Keiding (Kopenhagen); Masahiro Kuroda Kurasiki (Science and Art Tokio); Steffen Lauritzen (Aalborg ); Thomas Richardson (Warwick); Nanny Wermuth (Mannheim).
Paul Biemer (North Carolina); Raymond A. Carroll (Texas A&M); Montezuma Dumangane (Bristol); Jouni Kuha (Penn State); Bani K. Malik (Texas A&M); Geert Molenberghs (Limburg); Joseph Schafer (Penn State); Burkhard Schaffrin (Columbus/Ohio); Chris Skinner (Southhampton); Silvelyn Zwanzig (Hamburg).
Murray Aitkin (Newcastle); Dankmar Böhning, (Berlin); Jim Booth (Gainesville, Florida); David Clayton (Cambridge); Geert Molenberghs (Limburg) ; Geert Verbeke (Leuven)
Paul Albert (Maryland), Per Kragh Andersen (Kopenhagen), Wiji Arampulam (Warwick), John Bailer (Miami), Luc Bauwens (Louvain), Peter Brockwell (Colorado), Phil Dawid (London), Vanessa Didelez (London), Jürgen Franke (Kaiserslautern), Sylvia Frühwirth-Schnatter (Wien), J.C. van Houwelingen (Leiden), Tony Morton-Jones (Manchester), John Petkau (Vancouver), Mike Smith (Sydney), Werner Vach (Odense), Jian-Feng Yao (Rennes).
Adrian Bowman (Glasgow ); Theo Gasser (Zürich ); Peter Green (Bristol); Jim Ramsay (Canada); Matt Wand (Boston).
F. Coolen (Durham), G. de Cooman (Ghent), F. Kozine (Roshilde, DK), H. Rieder (Bayreuth), M. Zaffalon (Lugano.
David Balding (London); Carsten Wiuf (Boston); Gesine Reinert (Oxford); Paul Eilers (Leiden); Arndt von Haeseler (Düsseldorf); Nuala Sheehan (Leicester); Chris Holmes (London); Michael Newton (Madison); Terry Speed (Berkeley).
Anthony Davison (Swiss Federal Institute of Technology ); Arnoldo Frigessi (Norwegian Computing Center); Jeffrey Hart (Texas A&M); Brian Ripley (Oxford); Antony Unwin (Augsburg).
Ole Barndorff-Nielsen (Aarhus); Peter Brockwell (Colorado); Petros Dellaportas (Athen); Joachim Grammig (St. Gallen); Friedrich Hubalek (Vienna); Jan Kallsen (Freiburg); Christoph Kühn (München); Stefan Mittnik (Kiel / München); Winfried Pohlmeier (Konstanz); Martin Schlather (Bayreuth); Neil Shephard (Oxford); Niklas Wagner (München).
Yasuo Amemiya (New York); Badi Baltagi (Texas); Steve Bond (Oxford); Peter Bühlmann (Zürich); Bernd Fitzenberger (Mannheim); Harald Hruschka (Regensburg); Hans-Joachim Lenz (Berlin); Costas Meghir (London); Germano Mwabu (Nairobi); Regina Riphahn (Basel); Gerd Ronning (Tübingen); Christoph Schmidt (Essen); Paul Schultz (Yale); Wilfried Seidel (Hamburg); Tom Wansbeek (Groningen); Rainer Winkelmann (Zürich).
Aalen, O. (Oslo); Albert, P. (Maryland); Beibel, M. (Freiburg); Cheng, C. (Taipeh); Davies, L. (Essen); Edwards, D. (Kopenhagen); Fokianos, K. (Columbus/Ohio); Gemperli, A. (Bern); Heuer, C. (Heidelberg); Hruschka, H. (Regensburg); Lang, J. (Iowa); Nikitin, Y. (St. Petersburg); Rieder, H. (Bayreuth); Rödel, E. (Berlin); Santner, T. (Columbus/Ohio); Schaffrin, B. (Columbus/Ohio); Sheehan, N. (Loughborough); Smith, M. (Sydney); Spokoiny, V. (Berlin); Stasinopoulos, D. (London); Timmer, J. (Freiburg); Wellner, J. (Washington); Wolters, J. (Berlin); Ziegler, A. (Marburg).
Berliner, M. (Columbus/Ohio); Briegel, T. und Tresp, V. (München); Fokianos, K. (Nikosia); Gamerman, D. (Rio de Janeiro); Gasmi, S. (Magdeburg); Gelfand, A. (Connecticut); Giudici, P. (Pavia); Hart, J. (Texas A&M); Hübner, U. und Herrmann, P. (Mannheim); Hujer, R. (Frankfurt); Keiding, N. (Kopenhagen); Kohn, R. (Sydney); Kukush, O. (Kiew); Mittag, H.-J. (Hagen); Müller, H.-G. (Berkeley); Pohlmeier, W. (Konstanz); Prigarin, S. (Novosibirsk); Rammelt, P. (Berlin); Richardson, S. (Villejuif, Frankreich); Santner, T.J. (Columbus/Ohio); Shalabh, (Chandigarh/Indien); Shao, J. (Wisconsin); Srivastava, V. (Lucknow/Indien); Stemann, D. (Hagen); Viertl, R. (Wien); Wermuth, N. (Mannheim); Zwanzig, S. (Hamburg).
Booth, J. (Gainsville, Florida); Cressie, N. (Columbus/Ohio); Friedl, H. (Graz); Frühwirth-S. S. (Wien); Hobart, J. (Florida, Gainsville); Marx, B. (Baton Rouge); Müller, P. (North Carolina); Opsomer, J. (Iowa); Rue, H. (Trondheim); Schaffrin, B. (Columbus/Ohio); Shalabh, (Chandigarh/Indien); Wilrich, T. (Berlin).
Bartels, R. (Sydney); Becker, C. (Dortmund); Brännäs, K. (Umea); Coolen, F. (Durham); Dannegger, F. (München); di Serio, C. (San Raffaele, Italien); Gelfand, A. (Connecticut/USA); Gerhard, F. Nuffield College, (Oxford); Göttlein, A. (München); Heikkinen, J. (Helsinki); Kauermann, G. (Glasgow); Knorr-Held, L. (London); Kukush, A. (Kiew); Smith, Michael, (Sydney); Smith, Murray, (Sydney); Odejar, M.A. (Los Banos, Philippines); Ranta, J. (Helsinki); Schaffrin, B. (Columbus/Ohio); Shafer, G. (Newark); Shklyar, S. (Kiew).
Adebayo, S. (Ilorin, Nigeria); Augustin, N. (Freiburg); Musio, M. (Freiburg); Berger, U. (Glasgow); Coolen F. (Durham), Chung, C.F. (Taipeh); de Cooman, G. (Ghent); Fokianos, K. (Nikosia); Friedl, H. (Graz); Gamerman, D. (Rio de Janeiro); Gieger, C. (Heidelberg); Grammig, J. (St. Gallen); Hafner R. (München); Hart, J. (Texas A&M ); Hipp, C. (Karlsruhe); Knorr-Held, L. (Lancaster); Kukush, A. (Kiew); van der Linde, A. (Bremen); Mira, A. (Varese/Italien); Muliere, P. (Mailand); Odejar, A. (Los Banos, Philippines); Schaffrin, B. (Columbus/Ohio); Shalabh, (Chandigarh/Indien); Smith, Mike, (Sydney); Utkin, L.V. (St. Petersburg), Zaffalon, M. (Lugano).
Alonso, A. (Limburg); Belenkiy, S. (Rußland); Bender, R. (Bielefeld); Carroll, R. (Texas A&M); Cheng, C. (Taipeh); Coolen, F. (Durham); Davidov, O. (Haifa, Israel); Höhle, M. (Aalborg); Kauermann, G. (Bielefeld); Lenz-Tönjes, R. (Bielefeld); Mackerras, D. (Alice Springs, Australia); Mansmann, U. (Heidelberg); Marx, B. (Baton Rouge); Molenberghs, G. (Limburg); O'Neill, P. (Nottingham); Santner, T. (Columbus/Ohio); Seeber, G. (Innsbruck); Shalabh, (Chandigarh/Indien); Wood, S. (Glasgow); Vardeman, S. (Iowa State ); Vontheim, R. (Tübingen); Zucchini, W. (Göttingen).
Adebayo, S. (2003). Semiparametric Bayesian Regression for Multivariate Responses. Hieronymus, München.
Adebayo, S. (2003a). Bayesian Geoadditive Modelling of Breastfeeding Initiation in Nigeria. To appear in: Journal of Applied Econometrics.
Adebayo, S. (2003b). Modelling Childhood Malnutrition in Zambia: an Adaptive Bayesian Splines Approach. To appear in: Statistical Methods and Applications, Journal of the Italian Statistical Society.
Auer, D.P., Pütz, B., Gössl, C., Elbel, G-K., Gasser, Th., Dichgans M. (2000): Differential lesion pattern in CADASIL and spontaneous subcortical arteriosclerotic encephalopathy: a MRI study using statistical parametric group comparison. To appear in: Radiology.
Augustin, T. (1998): Optimale Tests bei Intervallwahrscheinlichkeit. Vandenhoeck und Ruprecht, Göttingen.
Augustin, T. (1999a): Globally least favorable pairs and Neyman-Pearson testing under interval probability. In: G. de Cooman, F.G. Cozman, S. Moral and P. Walley (eds.): ISIPTA'99: Proceedings of the First International Symposium on Imprecise Probabilities and their Applications. Gent, 15-24.
Augustin, T. (1999b): On data-based checking of hypotheses in the presence of uncertain knowledge. In: W. Gaul and H. Locarek-Junge, H. (eds.): Classification in the Information Age. Springer. Heidelberg, 127-135.
Augustin, T. (2001). On decision making under ambiguous prior and sampling information. In: G. de Cooman, T. Fine, S. Moral, T. Seidenfeld (eds.): ISIPTA 01: Proceedings of the Second International Symposium on Imprecise Probabilities and their Applications. Cornell University, Ithaca (N.Y.), Shaker, Maastricht, 9-16.
Augustin, T. (2002a). Neyman-Pearson testing under interval probability by globally least favorable pairs - Reviewing Huber-Strassen theory and extending it to general interval probability. Journal of Statistical Planning and Inference 105, 149-173.
Augustin, T. (2002b): Expected utility within a generalized concept of probability – A comprehensive framework for decision making under ambiguity. Statistical Papers 43, 5-22.
Augustin, T. (2002c): Some basic results on the extension of quasi-likelihood based measurement error correction to multivariate and flexible structural models. In: W. Gaul and G. Ritter (eds.): Classification, Automation, and New Media. Springer, Heidelberg, 29-36.
Augustin, T. (2003a). An exact corrected log-likelihood function for Cox's proportional hazards model. To appear in: Scandinavian Journal of Statistics.
Augustin, T. (2003b). On the suboptimality of robust Bayesian procedures from the decision theoretic point of view. To appear in: J.M. Bernard, T. Seidenfeld, M. Zaffalon (Hg.): ISIPTA 03: Proceedings of the Third International Symposium on Imprecise Probabilities and their Applications, Lugano. Carleton Scientific, Waterloo, 31-45.
Augustin, T. (2003c): Optimal decisions under complex uncertainty – basic notions and a general algorithm for data-based decision making with partial prior knowledge described by interval probability. To appear in: ZAMM – Zeitschrift für Angewandte Mathematik und Mechanik.
Augustin, T. and Coolen, F. (2003): Nonparametric predictive inference and interval probability. To appear in: Journal of Statistical Planning and Inference.
Augustin, T. and Schwarz, R. (2002). Cox's proportional hazards model under covariate measurement error - A review and comparison of methods. In: Van Huffel, S., Lemmerling, P. (eds.): Total Least Squares and Errors-in-Variables Modeling: Analysis, Algorithms and Applications. Kluwer, Dordrecht. 175-184.
Augustin, T. and Wolff, J. (2003). A bias analysis of Weibull models ander heaped data. To appear in: Statistical Papers.
Becker, U., Fahrmeir, L. (2001). Bump Hunting for Risk: a New Data Mining Tool and its Applications. Computational Statistics, 16, 373-386.
Bender, S., Fahrmeir, L., Lang, S. (1998). Semiparametric Bayesian Analysis of Discrete Time Duration Data. In: B. Marx, H. Friedl (Hrsg.): Statistical Modeling, Proceedings of the 13th International Workshop on Statistical Modeling, New Orleans, USA, 115-122.
Bender, S., Fahrmeir, L., Lang, S. (2000). Determinanten der Arbeitslosigkeitsdauer in Westdeutschland. In: F. Büchel, M. Diewald, P. Krause, A. Mertens, H. Solga (Hrsg.): Zwischen drinnen und draußen. Arbeitsmarktchancen und soziale Ausgrenzungen in Deutschland. Verlag Leske + Budrich, Leverkusen.
Biller, C. (2000a). Bayesianische Ansätze zur nonparametrischen Regression. Shaker Verlag, Aachen.
Biller, C. (2000b). Adaptive Bayesian Regression Splines in Semiparametric Generalized Linear Models. Journal of Computational and Graphical Statistics 9, 122-140.
Biller, C. (2000c). Discrete Duration Models combining Dynamic and Random Effects. Lifetime Data Analysis 6, 375-390.
Biller, C. (2001). Posterior mode estimation in dynamic generalized linear mixed models. Allgemeines Statistisches Archiv 85, 27-43.
Biller, C., Fahrmeir, L. (2001). Bayesian varying-coefficient models using adaptive regression splines. Statistical Modelling 2, 195-211.
Blauth, A., Pigeot, I. und Bry, F. (2000). Interactive analysis of high-dimensional association structures with graphical models. Metrika 51, 53-65.
Boulesteix, A., Tutz, G., Strimmer, K. (2003). A CART-based Approach to Discover Emerging Patterns in Microarray Data,Bioinformatics 19, 2465-2472.
Caputo, A., Heinicke, A., Pigeot, I. (1999). A graphical chain model derived from a model selection strategy for the sociologists graduates study. Biometrical Journal 41, 217-234.
Caputo, A., Foraita, R., Klasen, S. und Pigeot, I. (2003). Undernutrition in Benin – an analysis based on graphical models. Social Science and Medicine 56, 1677-1691.
Carvalho, M. S. and Knorr-Held, L. (2003). Modelling discrete time survival data with random slopes: evaluating hemodialysis centres. Statistics in Medicine, 22, 3543-3555.
Cheng, C.-L., Schneeweiß, H. (1998). The polynomial regression with errors in the variables. Journal of the Royal Statistical Society B 60, 189-199.
Cheng, C.-L., Schneeweiß, H., Thamerus, M. (2000). A small sample estimator for a polynomial regression with errors in the variables. Journal of the Royal Statistical Society B 62, 699-709.
Cheng, C.-L. and Schneeweiss, H. (2002). On the polynomial measurement error model. In: van Huffel, S. and Lemmerling, P. (eds.): Total Least Squares and Errors-in-Variables Modeling – Analysis, Algorithms and Applications. Kluwer, Dodrecht-Boston-London. 131-143.
Chiarella, C., Semmler, W., Mittnik, S., Zhu, P. (2002). Stock Market, Interest Rate and Output: A Model and Estimation for U.S. Time Series Data. Studies in Nonlinear Dynamics and Econometrics, 6, Issue 1.
Claessen, H., Mittnik, S. (2002). Forecasting Stock Market Volatility and the Informational Efficiency of the DAX-Index Options Market. European Journal of Finance, 8, 302-321.
Crook, A., Knorr-Held. L., Hemingway, H. (2003). Measuring spatial effects in time to event data: a case study using months from angiography to coronary artery bypass graft. Statistics in Medicine 22, 2943-2961.
Dannegger, F. (2000). Tree stability diagnostics and some remedies for instability. Statistics in Medicine 19, 475-491.
Diggle, P., Knorr-Held, L., Rowlingson, B., Su, T.-L., Hawtin, P., Bryant, T. (2003). Towards on-line spatial surveillance. In: Brookmeyer, R. and Stroup, D. (eds.): Monitoring the Health of Populations: Statistical Principles and Methods for Public Health Surveillance. Oxford University Press.
Didelez, V., Pigeot, I. (1998). Maximum likelihood estimation in graphical models with missing values. Biometrika 85, 960-966.
Didelez, V. (2002). ML– and semiparametric estimation in logistic models with incomplete covariate data. Statistica Neerlandica 56, 330-345.
Didelez, V. (2003). Comments on graphical models for time series. In: Green, P., N. Hjort and S. Richardson (eds.): Highly structured stochastic systems. University Press, Oxford.
Didelez, V., Pigeot, I., Dean, K. and Wister, A. (2002). A comparative analysis of graphical interaction and logistic regression modelling: self-care and coping with a chronic illness in later life. Biometrical Journal 44, 410-432.
Doganoglu, T. (2002). Network Competition and Access Charge Rules. The Manchester School, 70, 16-35.
Doganoglu, T. (2003). Dynamic Price Competition with Consumption Externalities. NETNOMICS: Economic Research and Electronic Networking, 5, 43-69.
Dreesman, J., Tutz G. (2001) Nonstationary conditional models for spatial data based on varying coefficients. Journal of the Royal Statistical Society D 50, 1-15.
Drummond, A., and K. Strimmer. 2001. PAL: An object-oriented programming library for molecular evolution and phylogenetics. Bioinformatics 17, 662-663.
Edlich, S., Kauermann, G., Tutz G. (1998) Smoothing ordinal data by semiparametric models. Proceedings of the 13th International Workshop on Statistical Modelling.
Einbeck, J., Kauermann, G. (2003). Online monitoring with local smoothing methods and adaptive ridging. Journal of Statistical Computation and Simulation 73, 913-929.
Einbeck, J. (2003). Multivariate Local fitting with General basis functions. Computational Statistics 18, 185-203.
Einbeck, J., de André C.D.S., Singer, J.M. (2002). Local smoothing with robustness against outlying predictors. To appear in: Environmetrics.
Fahrmeir, L., Klinger, A. (1998). A nonparametric multiplicative hazard model for event history analysis. Biometrika 85, 581-592.
Fahrmeir, L., Pigeot, I., Künstler, R., Tutz, G., Caputo, A. , Lang, S. (1999). Statistik - Aufgabenbuch. Springer Verlag
Fahrmeir, L., Gieger.C., Heumann, C. (1999). An Application of Isotonic Longitudinal Marginal Regresssion to Monitoring the Healing Process. Biometrics 55, 951-956.
Fahrmeir, L. (1999). State space models: a brief history and some recent developments. ISI Bulletin, 1, 183-186.
Fahrmeir, L., Künstler, R. (1999). Penalized Likelihood Smoothing in Robust State Space Models. Metrika 49, 173-191.
Fahrmeir, L. (2000). Discussion on the Paper by Durbin and Koopman. Journal of the Royal Statistical Society B, 62, 42-43.
Fahrmeir, L. Knorr-Held, L. (2000). Dynamic and semiparametric models. In: Schimek, M.G. (ed.): Smoothing and Regression: Approaches, Computation and Application. Chap. 18, 505-536, Wiley, New York.
Fahrmeir, L., Lang, S., (2001a). Bayesian Inference for Generalized Additive Mixed Models Based on Markov Random Field Priors. Journal of the Royal Statistical Society C (Applied Statistics), 50, 201-220.
Fahrmeir, L., Lang, S., (2001b). Bayesian Semiparametric Regression Analysis of Multicategorical Time-Space Data. Annals of the Institute of Statistical Mathematics, 53, 10-30.
Fahrmeir, L., Tutz, G. (2001). Multivariate Statistical Modelling Based on Generalized Linear Models (revised and enlarged 2nd edition). Springer Verlag, New York.
Fahrmeir, L., Mayer, J. (2001). Bayesian-type count data models with varying coefficients: estimation and testing in the presence of overdispersion. Applied Stochastic Models in Business and Industry 17, 165-179.
Fahrmeir, L., Gössl, C. (2002). Semiparametric Bayesian Models for Human Brain Mapping. Statistical Modeling 2, 235-250.
Fahrmeir, L., Gössl, C., Hennerfeind, A. (2002). Robust Spatial Smoothing in functional MRI. In: Schwaiger, M. and Opitz, O. (eds.): Exploratory Data Analysis in Empirical Research. 50-57, Springer Verlag, Berlin.
Fahrmeir, L., Henking, A., Hüls, R. (2002). Methoden zum Vergleich von Scoreverfahren. www.rsiknews.de 11.2002.
Fahrmeir, L., Lang, S., Spies, F. (2003). Generalized Geoadditive Models for Insurance Claims Data. Blätter der Deutschen Gesellschaft für Versicherungsmathematik, 26, 7-23.
Fahrmeir, L., Lang, S., Wolff, J., Bender, S. (2003). Semiparametric Bayesian Time-Space Analysis of Unemployment Duration. Allgemeines Statistisches Archiv, 87, 281-307.
Fried, R., Didelez, V. (2003). Decomposability and selection of graphical models for time series. Biometrika 90, 251-267.
Friedl, H., Kauermann, G., (2000) Standard Errors for EM Estimates in Variance Component Models. Biometrics, 56, 761-767.
Giudici, P., Knorr-Held, L., Raßer, G. (2000). Modelling categorical covariates in Bayesian disease mapping by partition structures. Statistics in Medicine 19, 2579-2593.
Gössl, C., Auer, D. P., Fahrmeir, L. (2000). Dynamic models in fMRI. Magnetic Resonance in Medicine 43, 72–81.
Gössl, C., Fahrmeir, L. (2001). Simultaneous Bayesian Modeling of spatial effects and the HRF in fMRI. Proc. 10th Int. Symp. Appl. Stoch. Models and Data Analysis, 486-491.
Gössl C., Auer D.P., Fahrmeir L. (2001). Bayesian Spatiotemporal Inference in Functional Magnetic Resonance Imaging. Biometrics 57, 554-562.
Gössl C., Fahrmeir, L., Auer D.P. (2001). Bayesian Modeling of the Hemodynamic Response Function in BOLD fMRI. Neuroimage 14, 140-148.
Gössl, C., Küchenhoff, H. (2001). Bayesian analysis of logistic regression with an unknown change point and covariate measurement error. Statistics in Medicine 20, 3109-3121.
Gössl, C. (2002). Bayesian Models in Functional MRI: Approaches for Human Brain Mapping. Shaker, Aachen.
Gössl, C., Fahrmeir, L., Pütz, B., Auer, L.M., Auer, D.P. (2002). Fiber tracking from DTI using linear state space models: Detectability of the pyramidal tract. Neuroimage 16, 378-388.
Hansell, A., Knorr-Held, L., Schmid V. (2002). Mortality projections for obstructive lung disease in England & Wales. Thorax, 57, 151.
Hansell, A., Knorr-Held, L., Best. N, Schmid, V., Aylin, P. (2003). COPD mortality trends 1950-1999 in England & Wales – Did the 1956 Clean Air Act make a detectable difference? Epidemiology, 14 (5), 55.
Heid I.M., Gerken M, Wellmann J, Küchenhoff H, Kreienbrock L, Wichmann HE. (2002a). On the potential of measurement error to induce differential bias between groups: an example from radon epidemiology. Statistics in Medicine 21, 32613278.
Heid I.M., Schaffrath Rosario A, Kreienbrock L, Küchenhoff H, Wichmann HE. (2003). The impact of measurement error on studies on lung cancer and residential radon exposure in Germany. To appear in: J Tox Env Health.
Heinicke, A. (1999). Statistische Modellierung und Erfassung multivariater Assoziationsstrukturen anhand von KS-Verteilungen. Logos Verlag, Berlin.
Held, L., Natario, I., Fenton, S., Rue, H., Becker, N. (2003). A Joint Spatial Analysis of Cancer Mortality Data from Anatomical Sites with Related Risk Factors. To appear in: Statistical Methods in Medical Research.
Held, L. (2004). Simultaneous posterior probability statements from Monte Carlo output. To appear in: Journal of Computational and Graphical Statistics.
Heumann, C. (1998). Likelihoodbasierte marginale Regressionsmodelle für korrelierte kategoriale Daten. Dissertation. Peter Lang, Europäischer Verlag der Wissenschaften, Frankfurt am Main.
Höhle, M (2002). Decision support for pneumonia management in pig production. Proceedings of the First European Workshop on Sequential Decisions under Uncertainty in Agriculture and Natural Ressources, Toulouse, France, 51-56.
Kandala, N.B., Lang, S., Klasen, S., (2001). Semiparametric Analysis of Childhood Undernutrition in Developing Countries. In: George, E. (ed.): Monographs of Official Statistics; Bayesian methods with applications to science, policy and official statistics, 263-272.
Kandala, N.B., Lang, S., Klasen, S., Fahrmeir, L., (2001). Semiparametric Analysis of the Socio-Demographic and Spatial Determinants of Undernutrition in Two African Countries. Research in Official Statistics, 1, 81-100.
Kandala, N.-B. (2002). Spatial modelling of Socio-Economic and Demographic Determinants of Childhood Undernutrition and Mortality in Africa. Shaker, Aachen.
Kauermann, G., Müller, M., Carroll, R.J. (1998). The efficiency of bias-corrected estimators for nonparametric kernel estimation based on local estimation equations. Statistics & Probability Letters 37, 41-47.
Kauermann, G. ,Tutz, G. (1999) On Model Diagnostics and Bootstrapping in Varying Coefficient Models. Biometrika 86, 119-128.
Kauermann, G. (2000) Modelling Longitudinal Data with Ordinal Response by Varying Coefficients. Biometrics 56, 692-698.
Kauermann, G., Tutz, G. (2000) Local likelihood estimation in varying-coefficient models including additive bias correction. Journal of Nonparametric Statistics 12, 343-371.
Kauermann, G., Tutz, G. (2001) Testing generalized and semiparametric models against smooth alternatives. Journal of the Royal Society B, 63, 147-166
Kauermann, G. and Küchenhoff, H. (2003). Modelling Data from Inside of Earth: Local Smoothing of Mean and Dispersion Structure in Deep Drill Data. Statistical Modelling 3. 43-65.
Kauermann, G., Opsomer, J. (2003) Local Likelihood Estimation in Generalized Additive Models. Scandinavian Journal of Statistics, 30, 317-337.
Kauermann, G., Tutz G. (2003). Semiparametric Modelling of Ordinal Data. Journal of Computational and Graphical Statistics, 12, 176-196
Kaufmann, C., Pütz, B., Gössl, C., Auer D. P. (2000). Sex differences in a frequency graded visual stimulation paradigm for fMRI. Proceedings 6th Int. Conference on Functional Mapping of the Human Brain –HBM2000..
Kaufmann C., Pütz B., Gössl C., Elbel G. K., Auer D. P. (1999). Establishing a temporal demand and burden task for fMRI. NeuroImage 6, 130.
Klebanov, L.B., Mittnik, S., Rachev, S.T., Volkovich, V.E. (2000). A New Representation for the Characteristic Function of Strictly Geo-stable Vectors. Journal of Applied Probability, 37. 1137-1142.
Kleindienst, N., W. Greil, B. Rüger, H.J.-Möller (1999): The prophylactic efficacy of lithium - transient or persistent? Eur Arch Psychiatry Clin Neurosci, 229, Rothenburg.
Klinger, A. (1998). Hochdimensionale Generalisierte Lineare Modelle. Dissertation. Shaker Verlag, Aachen.
Klinger, A., Dannegger, F., Ulm, K. (2000). Identifying and modelling prognostic factors with censored data. Statistics in Medicine 19, 601-615.
Klinger, A., (2001). Inference in high dimensional generalized linear models based on soft thresholding. Journal of the Royal Statistical Society B, 63, 377-392.
Knorr-Held, L. (1999). Conditional prior proposals in dynamic models. Scandinavian Journal of Statistics 26, 129–144.
Knorr-Held, L. (2000). Bayesian modelling of inseparable space-time variation in disease risk. Statistics in Medicine, 19, 2555-2567.
Knorr-Held, L. (2003). Some remarks on Gaussian Markov random field models for disease mapping. In: Green, P., Hjort, N., Richardson, S. (eds.): Highly Structured Stochastic Systems. Oxford University Press.
Knorr-Held, L., Becker, N. (2000). Bayesian modelling of spatial heterogeneity in disease maps with application to German cancer mortality data. Allgemeines Statistisches Archiv 84, 121-140.
Knorr-Held, L., Besag, J.E. (1998). Modelling risk from a disease in time and space. Statistics in Medicine 17, 2045–2060.
Knorr-Held, L., Rainer, E. (2001). Prognosis of lung cancer mortality in West Germany: A case study in Bayesian prediction. Biostatistics 2, 109-129.
Knorr-Held, L., Richardson, S. (2003). A hierarchical model for space-time surveillance data on meningococcal disease incidence. Journal of the Royal Statistical Society C (Applied Statistics) 52, 169-183.
Knorr-Held, L, Rainer, E. (2001). Projections of lung cancer mortality in West Germany: a case study in Baysian prediction. Biostatistics 2, 109-129.
Knorr-Held, L., Raßer, G. (2000). Bayesian detection of clusters and discontinuities in disease maps. Biometrics 56, 13–21.
Knorr-Held, L., Raßer, G., Becker, N. (2002). Disease mapping of stage-specific cancer incidence data. Biometrics 58, 492-501.
Knorr-Held, L., Rue, H. (2002). On block updating in Markov random field models for disease mapping. Scandinavian Journal of Statistics 29, 597-614.
Knüsel, L. (1998). On the Accuracy of the Statistical Distributions in Microsoft Excel 97. Computational Statistics and Data Analysis 26, 375-377.
Knüsel, L. (2002). On the Reliability of Microsoft Excel XP for Statistical Purposes. Computational Statistics and Data Analysis 39, 109-110.
Knüsel, L. (2003a). On the Accuracy of Statistical Distributions in GAUSS 5.0. Computational Statistics and Data Analysis 42, 265-267.
Küchenhoff, H. and Pigeot I. (1998). Comparison of statistical software packages for finding a threshold limiting value in dose-response-relationships. Computational Statistics 13,117-140.
Kukush, A. and Maschke, E.O. (2003). The efficiency of adjusted least squares in the linear functional relationship. To appear in: Journal of Multivariate Analysis.
Kukush, A., Schneeweiß, H., Wolf, R. (2003). Three estimators for the Poisson regression model with measurement errors. To appear in: Statistical Papers.
Jerak, A., Lang, S. (2002). Locally Adaptive Function Estimation for Categorical Regression Models. In: Härdle, W., Rönz, B. (eds.): Proceedings in Computational Statistics (Compstat2002), 425-430, Physika Verlag, Heidelberg.
Lang, S., Adebayo, S., Fahrmeir, L., (2002). Bayesian Semiparametric Seemingly Unrelated Regression. In: Härdle, W., Rönz, B. (eds.): Proceedings in Computational Statistics, 195-200, Physika-Verlag, Heidelberg.
Lang, S., Fronk, E.M., Fahrmeir, L. (2002). Function Estimation with Locally Adaptive Dynamic Models. Computational Statistics 17, 479-500.
Lang, S., Kragler, P., Haybach, G., Fahrmeir, L. (2002). Bayesian space-time analysis of health insurance data. In: Schwaiger, M., Opitz, O. (eds): Exploratory Data Analysis in Empirical Research. 133-140, Springer Verlag, Berlin.
Lang, S., Brezger, A., (2003). Bayesian P-Splines. To appear in: Journal of Computational and Graphical Statistics.
Lang, S., Adebayo, S., Fahrmeir, L., Steiner, W., (2003). Bayesian Geoadditive Seemingly Unrelated Regression. Computational Statistics 18, 163-192.
Lang, S., Sunder, M., (2003). Nonparametric Regression with BayesX: A Flexible Estimation of Trends in Human Physical Stature in 19th Century America. Economics and Human Biology, 1, 77-90.
Leuzinger-Bohleber, M., U. Stuhr, B. Rüger, M. Beutel (2000): How to study the quality of psychoanalytic treatments? A combination of qualitative and quantitative approaches in a representative follow-up study. Paper presented at the Joseph Sandler Research Conference, London, March 4, 2000.
Leuzinger-Bohleber, M., U. Stuhr, M. Beutel, B. Rüger, H. Kächele (2001): Efectos a Largo Plazo del Psiconanálisis y de las Terapias Psicoanaliticas: un Estudio representativo de Seguimiento In: Florenzano, R., J.P. Jimènez, G. De la Parra, P. Riumalló, S. Valdivieso (eds.): Investigación en Psicoanálisis y Psicoterapia. IPA/SPR, Santiago de Chile, S.73-85.
Leuzinger-Bohleber, M. Stuhr, U., Rüger, B., Beutel, M. (2001a): Langzeitwirkungen von Psychoanalysen und Psychotherapien: Eine multiperspektivische, repräsentative Katamnesestudie. Psyche 55, 193-276.
Leuzinger-Bohleber, M. Stuhr, U., Rüger, B., Beutel, M. (2001b): Langzeitwirkungen von Psychoanalysen und Psychotherapien: Eine multiperspektivische, repräsentative Katamnesestudie. In: Bohleber, W., Drews, S. (eds.): Die Gegenwart der Psychoanalyse - die Psychoanalyse der Gegenwart, Klett-Cotta, Stuttgart, 567-590.
Leuzinger-Bohleber, M., U. Stuhr, B. Rüger, M. Beutel (2002): Forschen und Heilen Berichte und Ergebnisse psychoanalytischer Langzeittherapien. To appear at Kohlhammer Verlag, Stuttgart 2002.
Mittnik, S., Neumann, T. (2001). Dynamic Effects of Public Investment. Vector Autoregresssive Evidence from Six Industrialized Countries. Empirical Economics 26, 429-446.
Mittnik, S., Neumann, T. (2003). Time Series Evidence on the Non-Linearity Hypothesis for Public Spending. Economic Inquiry 41, 565-573.
Mittnik, S., Paolella, M.S. (2000). Conditional Density and Value-at-Risk Prediction of AsianCurrency Exchange Rates. Journal of Forecasting 19, 313-333.
Mittnik, S., Paolella, M.S. (2003). Prediction of Financial Downside-Risk with Heavy-Tailed Conditional Distributions. In: Rachev, S.T. (ed.): Handbook of Heavy Tailed Distributions in Finance, Elsevier, North-Holland, 384-404.
Mittnik, S., Paolella, M.S., Rachev, S.T. (2000). Diagnosing and treating the fat tails in financial return data. Journal of Empirical Finance 7, 389-416.
Mittnik, S., Paolella, M.S., Rachev, S.T. (2002). Stationarity of Stable Power-GARCH Processes. Journal of Econometrics 106, 97-107.
Mittnik, S., Paulauskas, V., Rachev, S.T. (2001). Statistical Inference in Regression with Heavy-tailed Integrated Variables. Mathematical and Computer Modelling 34, 1145-1158. Special Issue on Stable Models in Finance.
Mittnik, S., Rachev, S.T., Samorodnitsky, G. (2001). The Distribution of Test Statistics for Outlier Detection in Heavy-Tailed Samples. Mathematical and Computer Modelling, 34, 1171-1183.
Mittnik, S., Rachev, S., Schwartz, E. (2002). Value-at-risk and Asset Allocation with Stable Return Distributions. Allgemeines Statistisches Archiv (Journal of the German Statistical Society) 86, 53-67.
Mittnik, S., Rieken, S. (2000) Put-Call Parity and the Informational Efficiency of the German DAX-Index Options Market. International Review of Financial Analysis 9, 259-279.
Mittnik, S., Rieken, S. (2000). Lower-boundary Violations and Market Efficiency: Evidence from the German DAX-index Options Market. Journal of Futures Markets 20, 406-424.
Natario, I., Knorr-Held, L. (2003). Non-parametric ecological regression and spatial variation. Biometrical Journal 45, 670-688.
Nittner, T. (2002a). The Additive Model with Missing Values in the Independent Variable – Theory and Simulation. To appear in: Computational Statistics.
Nittner, T. (2002b). Missing at Random (MAR) in Nonparametric Regression – A Simulation Experiment. To appear in: Journal of the Italian Statistical Society.
Nittner, T. (2003b). Fehlende Daten in additiven Modellen. Band 7 von Anwendungsorientierte Statistik. Peter Lang Europäischer Verlag der Wissenschaften, Frankfurt/Main.
Paradis, E., Claude, J., Strimmer, K. (2003). APE: an R package for analyses of phylogenetics and evolution. To appear in: Bioinformatics.
Pigeot, I., Caputo, A., Heinicke, A., Brüderl, J. (2000). The professional career of sociologists: a graphical chain model reflecting early influences and associations. Allgemeines Statistisches Archiv 84, 3-21.
Rachev, S.T., Mittnik, S., (2000). Stable Paretian Modeling in Finance, Chichester. John Wiley & Sons.
Rao, C. R., Srivastava, V. K. & Toutenburg, H. (1998). Pitman nearness comparisons of stein-type estmators for regression coefficients in replicated experiments. Statistical Papers 39, 61-74.
Rao, C. R., Toutenburg, H.(1999). Linear Models: Least Squares and Alternatives (second enlarged and extended edition). Springer, New York.
Rao C. R., Toutenburg, H. (2001): Linear Models: Least Squares and Alternatives, reprint of the 2nd ed. for the Far East., Springer Hong Kong.
Rüger, B. (1998): Kumulative Niveauprofil-Tests. Jahrbücher für Nationalökonomie und Statistik 217, 82-92.
Rüger, B. (1998a): Über statistische Methoden in der Psychotherapieforschung. In: Fäh-Barwinski, M., G. Fischer (eds.) Sinn und Unsinn in der Psychotherapie-Forschung, Gießen, 52-81.
Rüger, B. (1999): Test- und Schätztheorie I: Grundlagen. Oldenbourg, München-Wien.
Rüger, B. (2001): Statistisches Design und Repräsentativität der Psychoanalytischen Katamnese-Studie (DPV). In: Stuhr, U., Leuzinger-Bohleber, M., Beutel, M. (eds.): Langzeit- Psychotherapie, Kohlhammer, Stuttgart, 149-159.
Rüger, B. (2002a): Statistical Design and Representativeness of the DPV-follow-up study. In: Leuzinger-Bohleber, M., Target, M. (eds.): Outcomes of Psychoanalytic Treatment, Whurr, London-Philadelphia, 121- 129..
Rüger, B. (2002): Test- und Schätztheorie. Band II: Statistische Tests. Oldenbourg, München.
Salemi, M., K. Strimmer, W. W. Hall, M. Duffy, E. Delaporte, S. Mboup, M. Peeters, and A.-M. Vandamme. (2001). Dating the common ancestor of SIVcpz and HIV-1 group M and the origin of HIV-1 subtypes using a new method to uncover clock-like molecular evolution. FASEB J. 15, 276-278.
Sailer, K. (2001). Regulierungsbedarf in Netzwerken? Implikationen für die Internetökonomie. Die Weltwirtschaft, 4.
Schmidt, H. A., K. Strimmer, M. Vingron, A. von Haeseler. (2002). TREE-PUZZLE: Maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics 18, 502-504.
Schneeweiß, H. (1998a). Korrigierte Schätzgleichungen für allgemeine Regressionsmodelle mit Fehlern in den Variablen. Jahrbücher für Nationalökonomie und Statistik 217, 116-122.
Schubert, Ch. et al. (2001). The Influence of Daily Psychosocial Incidents and Associated Emotions on the Course of Biochemical Parameters in SLE: Experience Taken from Three “Integrative Single-case Studies”, Psychosomatic Medicine, 63, 170.
Shalabh, Toutenburg, H. (2003): Pseudo Minimax Linear and Mixed Regression Estimation of regression Coefficients When Prior Estimates are Available. Statistics and Probability Letters 63, 35-39.
Shao, J., Kübler, J., Pigeot, I. (2000). Consistency of the bootstrap procedure in individual bioequivalence. Biometrika 87, 573-585.
Simonoff, J., Tutz, G. (1999). Smoothing methods for discrete data. In: M. Schimek (ed.): Smoothing and Regression. Computation and Application, Wiley.
Smith, M., Pütz, B., Auer, D., Fahrmeir, L. (2003). Assessing Brain Activity through Spatial Bayesian Variable Selection. Neuroimage, 17, 802-815.
Spiess, M., Tutz, G. (2002). Alternative measures of the explanatory power of multivariate probit models with continuous or ordinal responses. Journal of Mathematical Sociology,
Stasinopoulos, D.M., Rigby, R.A., Fahrmeir, L. (2000). Modelling rental guide data using mean and dispersion additive models. The Statistician, 49, 479-493.
Staubach, C., Schmid, V., Knorr-Held, L., Ziler, M. (2002). A Bayesian model for spatial wildlife disease prevalence data. Preventive Veterinary Medicine 56(1), 75-87.
Strimmer, K., and V. Moulton. 2000. Likelihood analysis of phylogenetic networks using directed graphical models. Mol. Biol. Evol. 17, 875-881.
Strimmer, K., O. G. Pybus. (2001). Exploring the demographic history of DNA sequences using the generalized skyline plot. Mol. Biol. Evol. 18, 2298-2305.
Strimmer, K., C. Wiuf, V. Moulton (2001). Recombination analysis using directed graphical models. Mol. Biol. Evol. 18, 97-99.
Strimmer, K., A. Rambaut (2002). Inferring confidence sets of possibly misspecified gene trees. Proc. R. Soc. Lond. B 269, 137-142.
Strimmer, K. 2003. Modeling gene expression measurement error: a quasi-likelihood approach. BMC Bioinformatics 4, 10.
Strimmer, K., K. Forslund, B. Holland, and V. Moulton. 2003. A novel exploratory method for visual recombination detection. Genome Biology 4, R33.
Thamerus, M. (1998a). Nichtlineare Regressionsmodelle mit heteroskedastischen Meßfehlern. Dissertation. Logos, Berlin.
Thamerus, M. (2003). Fitting a mixture distribution to a variable subject to heteroscedastic measurement errors. Computational Statistics 18, 1-17.
Toutenburg, H., Fieger, A., Kastner, C. (1998): Deskriptive Statistik. Prentice Hall, München.
Toutenburg, H. & Srivastava, V. K. (1998). Estimation of ratio of population means in survey sampling when some observations are missing. Metrika 48, 177-187
Toutenburg, H., Fieger, A., Srivastava, V. K. (1999): Weighted modified first order regression procedures for estimation in linear models with missing X-observations, Statistical Papers 40, 351-361.
Toutenburg, H., Srivastava, V. K. (1999): Improving the estimation of Coefficients in Linear Regression Models with some Missing Observations on some Explanatory Variables. Journal of the Italian Statistical Society 8(2), 191-204.
Toutenburg H. (2000): Deskriptive Statistik. 3. Aufl., Springer-Verlag, Heidelberg.
Toutenburg, H. (2000): Induktive Statistik, 2. Aufl., Springer-Verlag, Heidelberg.
Toutenburg, H., Fieger, A., Heumann, C. (2000): Regression Modelling with Fixed Effects: Missing Values and Related Problems. In: Statistics for the 21st century, Methodologies for Applications of the Future, 161, 423-439.
Toutenburg, H., Shalabh (2000). Improved predictions in linear regression models with stochastic linear constraints. Biometrical Journal 42, 71-86.
Toutenburg, H., Shalabh (2001). Use of minimum risk approach in the estimation of regression models with missing observations. Metrika 54, 247-259.
Toutenburg, H., Srivastava, V. K. (2001): Estimation of coefficients in linear regression models with some missing observations on an explanatory variable: Effect of departure from normality. Far East Journal of Theoretical Statistics 5, 17-36.
Toutenburg, H. (2002): Lineare Modelle, 2. Aufl., Heidelberg.
Toutenburg, H. (2002): Statistical Analysis of Designed Experiments (2nd ed.). Springer-Verlag, New York.
Toutenburg, H., Fieger, A., Schaffrin, B. (2002): Approximate Confidence Regions for Minimax-Linear Estimators. In: A. Ullah, A. T. K. Wan, A. Chaturvedi (eds.), Handbook of Applied Econometrics and Statistical Inference, Chap. 2, 27-44, Marcel Dekker, New York.
Toutenburg, H., Heumann, C., Nittner, T., Scheid, S. (2002). Parametric and Nonparametric Regression with missing X’s - A Review. Journal of the Iranian Statistical Society, 1, Nos. 1-2, 79-110.
Toutenburg, H., Nittner, T. (2002). Linear Regression Models with Incomplete Categorical Covariates. Computational Statistics 17 , 215-232.
Toutenburg, H., Shalabh (2002): Prediction of response values in linear regression models from replicated experiments. Statistical Papers 43, 423-433.
Toutenburg, H., Srivastava, V. K. (2002): A Revisit to the Application of Weighted Mixed Regression Estimation in Linear Models with Missing Data. American Journal of Mathematical and Management Sciences 22(384), 281-302.
Toutenburg, H., Srivastava, V. K., Schaffrin, B., Heumann, C. (2003): Efficiency Properties of Weighted Mixed Regression Estimators. Metron 61(1), 91-103.
Toutenburg, H., Shalabh (2003): Estimation of regression models with equi-correlated responses when some observations on the response variable are missing. Statistical Papers 44, 217-232.
Tutz, G. (1998). Latent Trait Modelle für ordinale Beobachtungen- Die statistische und messtheoretische Analyse von Paarvergleichsdaten. Springer Verlag
Tutz, G., (1998). Time-varying coefficients for discrete panel data with application to bussiness tendency surveys. Jahrbücher für Nationalökonomie und Statistik, 217, 334-344.
Tutz, G. (2000) Die Analyse kategorialer Daten –eine anwendungsorientierte Einführung in Logit--Modellierung und kategoriale Regression. Oldenbourg-Verlag.
Tutz, G. (2003a). Generalized semiparametrically structured ordinal models. Biometrics, 59, 263 – 273.
Tutz, G. (2003b) Generalized semiparametrically structured mixed models. To appear in: Computational Statistics and Data Analysis.
Tutz, G., Kauermann, G. (1998). Locally weighted least squares in categorical varying-coefficient models. In. Econometrics in Theory and Practice (Ed. Galata, R., Küchenhoff, H.), Physica-Verlag. Heidelberg.
Tutz, G., Kauermann, G. (2003). Generalized linear random effects models with varying coefficients. Computational Statistics and Data Analysis 43, 13-28.
Tutz, G., Scholz, T. (2003a). Semiparametric modelling of multicategorical data. To appear in: Journal of Statistical Computation and Simulation.
Utkin, L.V., Augustin, T. (2003a). Decision making with imprecise second order probabilities. In: J.M. Bernard, T. Seidenfeld, M. Zaffalon (eds.): ISIPTA 03: Proceedings of the Third International Symposium on Imprecise Probabilities and their Applications, Lugano. Carleton Scientific, Waterloo, 547-561.
Wallner, A. (2002). Beiträge zur Theorie der Intervallwahrscheinlichkeit – Der Blick über Kolmogorov und Choquet hinaus. Dr. Kovac, Hamburg.
Wallner, A. (2003A). Bi-elastic neighbourhood models, J.M. Bernard, T. Seidenfeld, M. Zaffalon (eds.), Proceedings of the Third International Symposium on Imprecise Probabilities and Their Applications (ISIPTA), Lugano, Switzerland, Carleton Scientific, Waterloo, 593-607.
Wassmer, G., B. Rüger (1998): Profil-a-Tests. In: Encyclopedia of Statistical Sciences, 2, (ed.: Kotz S., Read C. and D. Banks), Canada.
Weichselberger, K. (2000). The theory of interval-probability as a unifying concept for uncertainty, International Journal of Approximate Reasoning 24, 149-179.
Weichselberger, K. (2001). Elementare Grundbegriffe einer allgemeineren Wahrscheinlichkeitsrechnung I. Intervallwahrscheinlichkeit als umfassendes Konzept. Physika-Verlag, Heidelberg. Cooperating T. Augustin and A. Wallner.
Weichselberger, K. (2002). Alternative probabilistic systems, UNESCO, Encyclopedia of Life Support Systems 6.2.1.4, to appear at http://www.eolss.net/.
Weichselberger, K., Augustin, T. (1998a): Zur Konstruktion von Bereichsschätzungen für nicht-monotone Funktionen von Parametern. Jahrbücher für Nationalökonomie und Statistik 217, 123–137.
Weichselberger, K., Augustin, T. (2003): On the competition and symbiosis of two concept of conditional interval probability. In: J.M. Bernard, T. Seidenfeld and M. Zaffalon (eds.): ISIPTA 03: Proceedings of the Third International Symposium on Imprecise Probabilities and their Applications, Lugano. Carleton Scientific, Waterloo, 608-629.
Wichert, S., Fokianos K., Strimmer K. (2003). Identifying periodically expressed transcripts in microarray time series data. To appear in: Bioinformatics.
Wolff, J. and Augustin, T. (2003). Heaping and its consequences for duration analysis – a simulation study. Allgemeines Statistisches Archiv – Journal of the German Statistical Association 87, 1-28.
Yasssouridis, A., Steiger, A., Klinger, A., Fahrmeir, L. (1999). Modelling and exploring human sleep with event history analysis. Journal of Sleep Research, 8, 25-36.
Ziegler, A., Blettner, M., Kastner, C., Chang-Claude J. (1998). Identifying influential families using regression diagnostics for generalized estimation equations. Genetic Epidemiology 15, 341-353.
Ziegler, A., Kastner, C., Blettner, M. (1998). The generalised estimating equations: An annotated bibliography. Biometrical Journal 40, 115-139.
Discussion Papers and Other Contributions
Most discussion papers are currently under review, under revision or revised for publication in refereed journals.
Adebayo, S., Fahrmeir, L. (2003). Analyzing child mortality in Nigeria with geoadditive survival models. Discussion Paper 303, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München. In Revision for Statistics in Medicine.
Augustin, T. (1999c). Quasi-likelihood based correction for measurement error in accelerated failure time models. In. Friedl, H., Berghold, A., Kauermann, G. (eds.). Statistical Modelling. (Proceedings of the 14th International Workshop on Statistical Modelling). Graz, 421-424.
Augustin, T. (1999d): Interval probability – towards a more comprehensive mathematical modeling of uncertainty. In: S. Maksay (ed.): Analele Facultatii de Inginerie din Hunedoara (Festschrift on the occasion of the anniversary of the Faculty of Engineering, University of Hunedoara (Romania), ISSN 1454-6531), 143-147.
Augustin, T. (2000). Correcting for measurement error in parametric duration models by quasilikelihood. Discussion Paper 157, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Augustin, T. (2002d). Survival Analysis Under Measurement Error. Habilitationsschrift. Ludwig-Maximilians-Universität München.
Augustin, T. (2003d): Generalized basic probability assignments. Discussion Paper 277, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München. Conditionally accepted in International Journal of General Systems.
Augustin, T., Döring, A., Rummel, D. (2003). Regression calibration for Cox regression ander heteroscedastic measurement error - Determining risk factors of cardiovascular diseases from error-prone nutritional replication data. Discussion Paper 345, Sonderforschungsbereich 386, Ludwig-Maximilians-University Munich.
Augustin, T., Pöhlmann, S. (2003): On robust sequential analysis – Kiefer-Weiss optimal testing under interval probability. Conditionally accepted in Journal of Statistical Planning and Inference.
Becker, U., Fahrmeir, L. (2000). Bump Hunting for Risk: a New Data Mining Tool. In: Data Analysis (ed. W. Gaul et al.), Springer, 395-404.
Bender, R., Augustin, T., Blettner, M. (2003). Simulating survival times for Cox regression models. Discussion Paper 338, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München. Under revision for Statistics in Medicine.
Bender, S., Kohlmann, A., Lang, S., (2003). Women, Work, and Motherhood: Changing Employment Penalties for Motherhood in West Germany after 1945. A Comparative Analysis of Cohorts Born in 1934-1971. Discussion Paper 309, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Berger, U., Fahrmeir, L. Klasen, S. (2002). Dynamic Modelling of Child Mortality in Developing Countries: Application for Zambia. Discussion Paper 299, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Boulesteix, A., Tutz, G. (2003) A Framework to Discover Emerging Patterns for Application in Microarray Data. Discussion Paper 313, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Brezger, A., (2003). Isotonic regression based on Bayesian P-splines. Discussion Paper 331, Sonderforschungsbereich 386 Ludwig-Maximilians-Universität München.
Brezger, A., Kneib, T., Lang, S., (2003a). BayesX Manual. Available at http:://www.stat.uni-muenchen.de/~lang.
Brezger, A., Kneib, T., Lang, S., (2003b). BayesX: Analysing Bayesian Semiparametric Regression Models. Discussion Paper 332, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Brezger, A., Lang, S., (2003). Generalized structured additive regression based on Bayesian P-splines. Discussion Paper 321, Sonderforschungsbereich Ludwig-Maximilians-Universität München.
Caputo, A. (1998a). Eine alternative Familie von Modellverteilungen für Kovarianz- und Konzentrationsgraphen. Dissertation, Universität München.
Caputo, A. (1998b). Graphical models with Koehler Symanowski distribution. Discussion Paper 104, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Caputo, A. (1998c). Comments on ML estimation in graphical models with KS distribution. In Marx, B., H. Friedl (Hrsg.) Proceedings of the 13th International Workshop on Statistical Modelling, New Orleans, 1998, 131-138.
Caputo, A. (1998d). Decomposition of ML estimation in graphical models with Koehler Symanowski distribution. Discussion Paper 105, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Caputo, A. (1998e). Some properties of the family of Koehler Symanowski distributions. Discussion Paper 103, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Caputo, A. (1999). Factorization of the cumulative distribution function in case of conditional independence. Discussion Paper 161, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Cheng, C.-L., Schneeweiß, H., Thamerus, M. (1998). A Small Sample Estimator for a polynomial regression with errors in the variables. Discussion Paper 113, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Denuit, M., Lang, S. (2003). Nonlife Ratemaking with Bayesian GAM's. Discussion Paper 304, Institute de Statistique, Universite Catholique de Louvain.
Didelez, V. (1999a). Local independence graphs for composable Markov processes. Discussion Paper 158, Sonderforschungsbereich 386, Ludwig Maximilians-Universität München.
Didelez, V. (1999b). Graphical modelling of local independence structures in multivariate event history analysis. Book of Abstracts of the 2nd European Conference on Highly Structured Stochastic Systems (Pavia, 14.-18.9.1999), 95-98.
Didelez, V. (1999c). Local independence - a new approach to graphical modelling of event history data. In Studeny, M.,, Matus, F. (Hrsg.) Book of Abstracts of the Workshop "Conditional independence structures and graphical models" (Toronto, 27.9.-1.10.1999), 25-26.
Didelez, V. (1999d). Graphische Modellierung von lokalen Abhängigkeitsstrukturen in der multivariaten Ereignisanalyse. In: Hennig, C., Pfeifer, D. (Hrsg.), Hamburger Beiträge zur Modellierung und Simulation, Heft 13, 50-52.
Didelez, V., Pigeot, I., Walter, P. (2000). Modifications of the Bonferroni-Holm procedure for a multiway ANOVA. Discussion Paper 186, Sonderforschungsbereich386, Ludwig Maximilians-Universität München.
Doganoglu, T. (2003b). Licensing with Consumer Heterogenity and Optimal Number of Licenses. Working Paper, CINE, Department of Statistics, University of Munich.
Doganoglu, T., Grzybowski, L. (2003). Dynamic Duopoly Competition with Switching Costs and Network Externalities. Working Paper, CINE, Department of Statistics, University of Munich.
Doganoglu, T., Mittnik, S. (2003). Portfolio Selection, Risk Assessment and Heavy Tails. Preprints, Center for Financial Studies Research Conference on New Directions in Financial Risk Management.
Doganoglu, T., Sailer, K. (2003). Special journal issue of NETNOMICS: Economic Research and Electronic Networking on The Microeconomics of the New Economy. Vol. 5.
Doganoglu, T., Wright, J. (2003). Multihoming and Compatibility. Working Paper, CINE, Department of Statistics, University of Munich.
Eberle, W., Toutenburg, H. (1999): Handling of missing values in statistical software packages for windows. Discussion Paper 170, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Einbeck, J. (2001). Local Fitting with General Basis Functions. Discussion Paper 256, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Einbeck, J. (2003) Local Smoothing Methods for the Analysis of Multivariate Complex Data Structures. Dissertation LMU München.
Einbeck, J., Tutz, G., Evers, L. (2003) Local principal curves. Discussion Paper 320, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Evers, L. (2003) Partitionierungsverfahren zur Dimensionsreduktion. Diplomarbeit LMU München.
Fahrmeir, L. (1999). Recent Advances in Semiparametric Bayesian Function Estimation. In: Gaul, W. et al. (eds.): Mathematische Methoden der Wirtschaftswissenschaften, Physica Verlag, 118-129.
Fahrmeir, L., Hennerfeind, A. (2003). Nonparametric Bayesian Hazard rate models based on penalized regression splines. Discussion Paper 361, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Fahrmeir, L., Gössl, C., Hennerfeind, A. (2003). Bayesian mapping of brain regions using compound Markov random field priors. Discussion Paper 317, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Fahrmeir, L., Kneib, T., Lang, S. (2003). Penalized additive regression for space-time data: a Bayesian perspective. Discussion Paper 305, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München. Revised für Statistica Sinica.
Fahrmeir, L., Osuna, L., (2003). Structured additive count data regression. Discussion Paper 334, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Fieger, A., Kastner,C., Heumann, C. (1998): WinMAREG Quick Start. SFB386-Discussion Paper 108, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Foraita, R. (2000). Determinanten der Unterernährung - ein Ländervergleich anhand graphischer Kettenmodelle. Diplomarbeit, Institut für Statistik, Ludwig Maximilians-Universität München.
Frank, M. (2000). Branchenorienterte Steuerung eines Kreditportfolios. Dissertation. Ludwig Maximilians-Universität München.
Frohwerk, S. (2003). An empirical study of the Sourceforge.net database. Working Paper, CINE, Department of Statistics, University of Munich.
Galindo, C., Kauermann, G. , Liang, H. and Carrol, R. (2000). Bootstrap Confidence Intervals for Local Likelihood, Local Estimating Equations and Varying Coefficients Models. Sonderforschungsbereich Discussion Paper 205
Gartner, H., Scheid, S. (2003): Multiple Imputation von fehlenden Werten mit Daten über Unterernährung und Kindersterblichkeit, Discussion Paper 322, Sonderforschungsbereich 386. Ludwig-Maximilians-Universität München.
Gehrmann, U., Hellriegel, B., Neiss, A., Fahrmeir, L. (2003). Analysis of the time to sustained progression in Multiple Sclerosis using generalised linear and additive models. Discussion Paper 354, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Gieger, Ch. (1999). Marginal Regression Models with Varying Coefficients for Correlated Ordinal Data. Discussion Paper 177, Sonderforschungsbereich 386, Ludwig Maximilians-Universität München, revised for Biometrics.
Grzybowski, L. (2003). The Competitiveness of Mobile Telecommunications Industry across the European Union. Working Paper, CINE, Department of Statistics, University of Munich.
Haas, M., Mittnik, S., Paolella, M.S. (2002): Mixed Normal Conditional Heteroskedasticity. Center for Financial Studies. Working Paper No. 10/2002.
Haas, M., Mittnik, S., Paolella, M.S. (2003). A New Approach to Markov-Switching GARCH Models. Preprints, Center for Financial Studies Research Conference on New Directions in Financial Risk Management.
Hartz, C., Mittnik, S., Paolella, M. (2003a). Forecasting Financial Time Series: Normal GARCH with Outliers or Heavy-Tailed Distribution Assumptions? Working Paper, Seminar for Financial Econometrics, Department of Statistics, University of Munich.
Hartz, C., Mittnik, S., Paolella, M. (2003b). Accurate Value-at-Risk Forecasting Based on the (good old) Normal-GARCH Model. Working Paper, Seminar for Financial Econometrics, Department of Statistics, University of Munich.
Heid, I.M., Küchenhoff, H., Miles, J., Kreienbrock, L., Wichmann, H.E. (2003) Two dimensions of measurement error: Classical and Berkson error in residential radon exposure assessment. In review since 02/2003.
Hennerfeind, A., Brezger, A., Fahrmeir, L. (2003). Geoadditive Survival Models. Discussion Paper 333, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Heumann, C. (1998). Likelihoodbasierte marginale Regressionsmodelle für korrelierte kategoriale Daten. Dissertation. Peter Lang, Frankfurt am Main.
Heumann, C., Fieger, A. und Kastner, C. (1998). C++ Utilities zur Implementierung statistischer Verfahren unter Berücksichtigung fehlender Werte. SFB386-Discussion Paper 109, Universität München.
Heumann, C. (2000). Intention-to-treat with drop-out. Discussion Paper 199, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Heumann, C. (2003). Monte Carlo methods for missing data in generalized linear and generalized linear mixed models. Habilitationsschrift, vorgelegt der Fakultät für Mathematik, Informatik und Statistik, Universität München.
Heumann, C. (2003). Intention-to-treat with drop-out, under revision für Biometrical Journal.
Höhle, M., Jørgensen, E. und O'Neill, P.D. (2003). Inference in disease transmission experiments using stochastic epidemic models. Invited for revision for Journal of the Royal Statistical Society, Series C (Applied Statistics).
Holmes, C. and Held, L. (2003). Bayesian auxiliary variable models for binary and polychotomous regression. Technical Report, Imperial College London.
Hübler, O., Mittnik, S., Schmid, W. (2002). Special journal issue of Allgemeines Statistisches Archiv onStatistisch-ökonometrische Risikoanalyse der Finanzmärkte, 86.
Jerak, A., Lang, S. (2003). Locally Adaptive Function Estimation for Binary Regression Models. Discussion Paper 310, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München. Eingereicht bei Biometric Journal.
Kandala, N.-B., Fahrmeir, L., Klasen, S. (2003). Geo-additive models of Childhood Undernutrition in three Sub Saharan African Countries. Discussion Paper 287, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Kastner, C., Fieger, A. und Heumann, C. (1999). New features in MAREG 0.2.0, SFB386-Discussion Paper 156, Universität München.
Kastner, A. (2000). Fehlende Werte bei korrelierten Beobachtungen. Dissertation. Universität München.
Kauermann, G. (2002) On a small sample Adjustment for the profile score function in semiparametric smoothing models. Sonderforschungsbereich 386, Discussion paper 162. Journal of Multivariate Analysis, 82, 471-485.
Kauermann, G., Carroll, R.J. (2000) The Sandwich Variance Estimator. Efficiency Properties and Coverage Probability of Confidence Intervals. Sonderforschungsbereich 386, Discussion Paper 189, to appear in JASA.
Kauermann, G., Tutz, G. (2000) Semiparametric Modeling of Ordinal Data. Sonderforschungsbereich 386, Discussion Paper 193.
Kauermann, G. (2001). Edge Preserving Smoothing by Local Mixture modelling. Sonderforschungsbereich 386, Discussion paper 255.
Kauermann, G., Tutz, G. (2001) Vanishing of Risk Factors for the Success and Survival of Newly Founded Companies. Sonderforschungsbereich 386, Discussion paper 242
Kauermann, G. and Opsomer , G., J.D. (2001). A fast method for implementing Generalized Cross-Validation in multi-dimensional nonparametric regression. To appear in Journal of Computational and Graphical Statistics.
Kauermann, G., Berger, U. (2003). A Smooth Test in Proportional Hazard Survival Models using Local Partial Likelihood Fitting. Lifetime Data Analysis (to appear).
Klein, I., Mittnik, S. (2002), (eds.).Contributions to Modern Econometrics: From Data Analysis to Economic Policy, Boston. Kluwer Academic Publishers.
Klodt, H. et al. (2003). Die neue Ökonomie: Erscheinungsformen, Ursachen und Auswirkungen. Eine Heinz Nixdorf Studie.
Knüsel, L. (2003b). On the Accuracy of Statistical Distributions in Microsoft Excel 2003. Submitted for publication in CSDA.
Knüsel, L. (2003c). Alternatives to the MCMC Method. Discussion paper 387, Sonderforschungsbereich 386, University of Munich.
Krause, R., Tutz, G. (2003a). Additive Modelling with Penalized Regression Splines and Genetic Algorithms. Sonderforschungsbereich 386, Discussion Paper 312.
Krause, R., Tutz, G. (2003b). Simultaneous selection of variables and smoothing parameter in additive models. (Eingereicht zum Tagungsband der Gesellschaft für Klassifikation).
Küchenhoff, H., Bender R., Langner, I., Lenz-Tönjes, R. (2003). Effect of Berkson measurement error on parameter estimates in Cox regression models. Discussion Paper 346, Sonderforschungsbereich 386, Universität München.
Küster, K., Mittnik, S., Paolella, M. (2003). Value-at-Risk Prediction: A Comparison of Alternative Strategies. Preprints, Center for Financial Studies Research Conference on New Directions in Financial Risk Management.
Kukush, A., Schneeweiß, H. (2000). A Comparison of Asymptotic Covariance Matrices of Adjusted Least Squares and Structural Least Squares in Error Ridden Polynomial Regression. Discussion Paper 218, Sonderforschungsbereich 386, Universität München.
Kukush, A., Schneeweiss, H. and Wolf, R. (2001a). Comparison of three estimators in a polynomial regression with measurement errors. SFBDiscussion Paper 233, Universität München.
Kukush, A., Schneeweiß, H., Wolf, R. (2002). Comparing Different Estimators in a Nonlinear Measurement Error Model. Discussion Paper 244, Sonderforschungsbereich 386, Universität München.
Lang, S., Brezger, A. (2000a). Bayesian P-Splines. Proceedings of IWSM, Bilbao.
Lang, S., Brezger, A. (2000b). Bayes X. Discussion Paper 187. Sonderforschungsbereich 386.
Lang, S., Fahrmeir, L., (2001). Bayesian generalized additive mixed models. A simulation study. Discussion paper 230, Sonderforschungsbereich 386 Ludwig-Maximilians-Universität München.
Lang, S., (2003). Structured additive estimation of mean and variance in Gaussian regression models. Discussion paper 330, Sonderforschungsbereich 386 Ludwig-Maximilians-Universität München. In Vorbereitung.
Lange, K. (2002). The Demand for Fixed-line Services of Households from Northern Germany. Working Paper, CINE, Department of Statistics, University of Munich.
Lange, K. (2003). The Demand for National Long-Distance Voice Telephony; Estimating Tariff Choices of Individual Households. Working Paper, CINE, Department of Statistics, University of Munich.
Lieske, N. (2000). Ansätze zur Intention-to-treat Analyse bei Drop-out. Diplomarbeit. Universität München.
Mittnik, S. (2001). Financial Modeling and Heavy Tails. In: W. Zhibao and T. Wansheng, eds. Proceedings of the IFAC Workshop on Computation in Economic, Financial and Engineering-Economic Systems, Tianjin, P.R. China, 437-445.
Mittnik, S., Rachev, S.T., Samorodnitsky, G. (2000). Testing for Structural Breaks in Time Series Regressions with Heavy-tailed Disturbances. Datamining and Computational Finance, Heidelberg. Physica-Verlag, 115-142.
Mittnik, S., Rachev, S.T. (2001). Special journal issue of Mathematical and Computer Modelling on Stable Non-Gaussian Models in Finance and Econometrics, 34.
Mittnik, S., Rachev S.T. (2002). Portfolio Selection in the Presence of Heavy-Tailed Asset Returns. In: I. Klein and S. Mittnik, eds.: Contributions to Modern Econometrics: From Data Analysis to Economic Policy, Boston. Kluwer Academic Publishers, 51-64.
Nittner, T. (1999). Fehlende Daten im klassischen linearen Regressionsmodell Eine Erweiterung existenter Verfahren auf diskrete Kovariablen. Diplomarbeit. Universität München.
Nittner, T. (2003a). Dissertation: Fehlende Daten in additiven Modellen.
Payer, T. and Küchenhoff, H. (2003). Modelling extreme wind speeds at a German weather station as basic input for a subsequent risk analysis for high speed trains. Journal of wind engineering and Industrial Aerodynamics, in press.
Pigorsch, C. (2003). Zinsprognose anhand der Zinsstruktur. Wirtschaft und Statistik, S. 548-564.
Pöhlmann, S., Augustin, T. (2001): A Kiefer-Weiss optimal sequential sign test – Some considerations on a bioequivalence problem from the viewpoint of quality management. In: J. Kunert and G. Trenkler (eds.): Mathematical Statistics with Applications in Biometry (Festschrift for Siegfried Schach), Josef Eul, Köln,179-188.
Prange, C. (2003a). Strompreisniveau – Vergleich EEX und Powernext. EMW Zeitschrift für Energie, Markt, Wettbewerb, 3: 71.
Prange, C. (2003b). Der heiße Sommer: Einfluss der Temperaturen auf den Strompreis. EMW Zeitschrift für Energie, Markt, Wettbewerb, 6 ,58-59.
Raßer, G. (2003). Clustering partition model for discrete structures. Dissertation.
Rummel, D. (2003). Glättung mit Basisfunktionen und randomisierten Gewichten. Diplomarbeit, Institut für Statistik, Universität München.
Sailer, K. (2001) Wettbewerb in elektronischen Märkten. In: Smekal, Ch. and H. Starbatty (Hrsg.). Veröffentlichung der Hanns Martin Schleyer-Stiftung, Band 58, Köln.
Schaffrin, B. & Toutenburg, H. (1998). The impact of missing values on the reliability measures in a linear model. Discussion Paper 125, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München, in Überarbeitung für Computational Statistics & Data Aanlysis.
Scheid, S. (2002). A Selection Model for Bivariate Normal Data, with a Flexible Nonparametric Missing Model and a Focus on Variance Estimates. Sonderforschungsbereich 386 Discussion Paper 285, München.
Schmid, V. (1999). Räumliche Erweiterungen von Bayesianischen Alter-Perioden-Kohortenmodellen. Diplomarbeit, Ludwig Maximilians-Universität München.
Schmid V. and Knorr-Held, L. (2003). Some remarks on Gaussian Markov random field models for disease mapping. In: Highly Structured Stochastic Systems, ed. By P. Green, N. Hjort, S. Richardson, Oxford University Press.
Schneeweiß, H. (1998b). Note on Two Estimators for the Polynomial Regression with Errors in the Variables. In: B. Marx, H. Friedl (Eds): Statistical Modeling, Proceedings of the 13th International Workshop on Statistical Modeling, New Orleans, Louisiana, USA, July 27-31, 141 -147.
Schneeweiß, H. (1999). Resolving the Ellsberg Paradox by Assuming that People Evaluate Repetitive Sampling. In: U. Leopold-Wildburger, G. Feichtinger, K.-P. Kistner (eds.): Modelling and Decisions in Economics. Essays in Honour of Franz Ferschl. Physica, Heidelberg, New York, 83-95.
Schneeweiß, H. (2003). Estimating the end point of a uniform distribution under measurement errors with known error variances. Discussion Paper 339, Sonderforschungsbereich 386, Universität München.
Schneeweiß, H., Cheng, C.-L., Wolf, R. (2002). On the bias of structural estimation methods in a polynomial regression with measurement error when the distribution of the latent covariate is misspecified. In: Klein, I., Mittnik, S. (ed.). Contributions in modern econometrics. From Data analysis to Economic Policy. In Honour of Gerd Hansen. Kluwer, Boston, 209-222.
Schneeweiß, H., Cheng, C.-L. (2003). Bias of a quasi score estimator of a measurement error model under misspecification of the regressor distribution. Discussion Paper 340, Sonderforschungsbereich 386, Universität München.
Schneeweiß, H., E.-O. Maschke, M. Pfannes (2001). The Reduced Form of Block Recursive Model. In: Ralph Friedmann, Lothar Knüppel, Helmut Lütkepohl (eds.): Econometric Studies. A Festschrift in Honour of Joachim Frohn.
Schneeweiß, H., Nittner, T. (2001). Estimating a polynomial regression with measurement errors in the structural and in the functional case – a comparison. In Saleh, M. (ed.): Data Analysis from Statistical Foundations, A Festschrift in Honour of the 75th Birthday of D.A.S. Fraser, Nova Sciences New York, 195-205.
Schubert, Ch. et al. (2003). The Impact of Emotionally Pleasing Incidents of Urine Cortisol and Urine Neopterin Dynamics in a Patient with Systemic Lupus Erythematosus. Working Paper. Department of Medical Psychology and Psychotherapy, University Hospital Innsbruck.
Seiffert-Klauss, V., Müller, J., Probst, R., Wilker, J., Höß, C., Luppa, P., Kastner, C. & Ulm, K. (2000). Bone metabolism during the perimenopausal transition: a prospective study. Discussion Paper 194, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München, eingereicht bei European Journal of Endocrynology.
Shklyar, S., Schneeweiss, H. (2002). A comparison of asymptotic covariance matrices of three consistent estimators in the Poisson regression model with measurement errors. Sonderforschungsbereich 386 Discussion Paper 283, Universität München. Submitted for publication.
Smith, M., Fahrmeir, L. (2002). Spatial Bayesian variable selection with application to functional magnetic resonance imaging. Submitted.
Srivastava, V. K. (1999): On the First Order regression Procedure of Estimation for Incomplete regression models. Discussion Paper 175, Sonderforschungsbereich 386. Ludwig-Maximilians-Universität München.
Storck, S., Kastner, C., Toutenburg, H. (2000): Longitudinal data with drop-outs: a comparison of pattern mixture models with complete case analysis. Discussion Paper 188, Sonderforschungsbereich 386. Ludwig-Maximilians-Universität München.
Thamerus, M. (1998b). Different nonlinear regression models with incorrectly observed covariates. In R. Galata, H. Küchenhoff (Hg.). Econometrics in Theory and Practice. Festschrift for Hans Schneeweiß. Physica, Heidelberg, New York, 31-44.
Toutenburg, H. & Shalabh (1998a). Use of minimum risk approach in the estimation of regression models with missing observation. Discussion Paper 118, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München, in Überarbeitung für Metrika.
Toutenburg, H. & Shalabh (1998b). Prediction of response values in linear regression models from replicated experiments. Discussion Paper 112, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München, eingereicht bei Statistical Papers.
Toutenburg, H. Fieger A. and Heumann C. (1998): Regression modelling with fixed effects - Missing values and related problems, SFB386-Discussion Paper 123, Universität München.
Toutenburg, H., Srivastava, V. K. (1998): Improving the Estimation of Incomplete Regression Models through Pilot Investigation and Replicated Studies, Discussion Paper 154, Sonderforschungsbereich 386. Ludwig-Maximilians-Universität München.
Toutenburg, H., Trenkler, G. (1998): Using Differences as a Device against Multicollinearity. In R. Galata, H. Küchenhoff, Herausgeber, Econometrics in Theory and Practice, Festschrift für Hans Schneeweiß, 131-136, Physica, Heidelberg.
Toutenburg, H., Shalabh (1998): Improved predictions in linear regression models with stochastic linear constraints, Discussion Paper 124, Sonderforschungsbereich 386. Ludwig-Maximilians-Universität München.
Toutenburg, H., Shalabh (1999): Estimation of regression Coefficients Subject to Exact Linear Restrictions when some Observations are Missing and Balanced Loss Function is Used, Discussion Paper 163, Sonderforschungsbereich 386. Ludwig-Maximilians-Universität München.
Toutenburg, H., Shalabh (1999): Estimation of regression Coefficients Subject to Exact Linear Restrictions when some Observations are Missing and Balanced Loss Function is Used, Discussion Paper 163, Sonderforschungsbereich 386. Ludwig-Maximilians-Universität München.
Toutenburg, H. & Srivastava, V. K. (1999). Amputation versus imputation of missing values through ratio method in sample surveys. Discussion Paper 155, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München, eingereicht bei Communications in Statistics, Theory & Methods.
Toutenburg, H., Fieger, A. (2000): Using diagnostic measures to detect non- MCAR processes in linear regression models with missing covariates. Discussion Paper 204, Sonderforschungsbereich 386. Ludwig-Maximilians-Universität München.
Toutenburg, H. and Nittner, T. (2000). The classical linear regression model with one incomplete binary variable. Discussion Paper 178, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München, eingereicht bei Computational Statistics.
Toutenburg, H. and Srivastava, V. K. (2000). Efficient estimation of population mean using incomplete survey data on study and auxiliary characteristics. Discussion Paper 179, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München, eingereicht bei Statistica.
Toutenburg, H., Shalabh (2001): A note on the comparison of minimax linear and mixed regression estimation of regression coefficients when prior estimates are available. Discussion Paper 238, Sonderforschungsbereich 386. Ludwig-Maximilians-Universität München.
Toutenburg, H., Shalabh (2001): Synthesizing the classical and inverse methods in linear calibration. Discussion Paper 252, Sonderforschungsbereich 386. Ludwig-Maximilians-Universität München.
Toutenburg, H., Shalabh (2001): Use of prior information in the form of interval constraints for the improved estimation of linear regression models with some missing responses. Discussion Paper 240, Sonderforschungsbereich 386. Ludwig-Maximilians-Universität München.
Toutenburg, H., Shalabh (2001): Estimation of linear models with missing data: The role of stochastic linear constraints. Discussion Paper 239, Sonderforschungsbereich 386. Ludwig-Maximilians-Universität München.
Trevisani, M., Causin, R., Montecchio, L., Kastner, C., & Heumann, C. (1999). Analysing the relationship between ectomycorrhizal infection and forest decline using marginal models. Discussion Paper 143, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München.
Tutz, G. (2001). Generalized semistructured mixed models. Sonderforschungsbereich 386, Discussion paper 251
Tutz, G., Binder, H. (2002). Flexible modelling of discrete failure time including time-varying smooth effects. Sonderforschungsbereich 386, Discussion Paper 294.
Tutz, G. (2002b). Generalized semistructured ordinal models. Sonderforschungsbereich 386, Discussion paper 250
Tutz, G., Edlich, S., Bäumer, C. (1999). Comparison between local estimates for multicategorial varying-coefficient models. Discussion Paper 151, Sonderforschungsbereich386, Universität München.
Tutz, G., Hechenbichler, K. (2003). Aggregating Classifiers With Ordinal Response Strucutre. Sonderforschungsbereich 386, Discussion Paper 359.
Tutz, G. (2003). Response smoothing estimators in binary regression. Sonderforschungsbereich 386, Discussion Paper 318.
Tutz, G., Scholz, T. (2003b) Ordinal regression modelling between proportional odds and non-proportional odds. Sonderforschungsbereich 386, Discussion Paper 304.
Uehlein, T. and Küchenhoff, H. (2003). Comparison of core inflation measures in the Euro area between 1995 and 2001. Allgemeines Statistisches Archiv 87, 405-424.
Utkin, L., Augustin, T. (2003b). Risk analysis on the basis of partial information about quantiles. In: S. Evgueni (ed.): Modelling and Analysis of Safety and Risk in Complex Systems (Proceedings MA SR 2003), St. Petersburg, 172-178.
Wallner, A. (2003b): Maximal number of vertices of polyhedra defined by 0/1-matrices: a basic result. Under revision for Discrete and Computational Geometry, Springer.
Weichselberger, K., Augustin, T. (1998b): Analysing Ellsberg's paradox by means of interval probability. In: R. Galata and H. Küchenhoff (eds.): Econometrics in Theory and Practice. Physica, Heidelberg, 291-304.
Ziegler, A., Kastner, C., Brunner D. & Blettner, M. (1999). Familial associations of lipid profiles: A generalised estimating equations approach. Discussion Paper 168, Sonderforschungsbereich 386, Ludwig-Maximilians-Universität München, submitted to Statistics in Medicine.
Zoppelli, L. (2003). Fehlende Beobachtungen im verallgemeinerten linearen Modell. Diplomarbeit, Institut für Statistik, Ludwig-Maximilians-Universität München.
