Informatique Pr Esent Ee Par : Statistical Learning and Regularisation for Regression Application to System Identiication and Time Series Modelling Apprentissage Statistique Et Regularisation Pour La R Egression
نویسندگان
چکیده
Application a l'identiication de syst emes et a la mod elisation des s eries chronologiques soutenue le 7 juillet 1997 devant le jury compos e de : Acknowledgements i Acknowledgements I am indebted to Professor Patrick Gallinari, who took me under his supervision in 1992, and followed me throughout this work. I would like him to know that his friendly support during these years were largely instrumental in the completion of this thesis. I would also like to thank Professor Lars Kai Hansen for welcoming me at the now-defunct Electronics Institute during the year 1995-1996, and later at the Institute of Mathematical Modelling. His enthusiastic comments and support had a major innuence on my work. I wish to express my gratitude to the members of the jury, who undertook to evaluate my work. Sylvie Thiria accepted to be the head of this jury; Anne Gu erin-Dugu e and G erard Govaert reviewed and refereed this thesis; Jan Larsen showed great interest in my work and accepted to y from Denmark to be part of this jury. I also wish to express my gratitude to all those who helped me through many fruitful discussions, or just by creating a friendly and challenging working atmosphere. The PhD students or ex-PhD students of the connectionist group of Laforia (Paris 6 University), and of the Digital Signal Processing group (Technical University of Denmark). Finally, special thanks are in order for my wife Kia. I don't know what I should thank her more for: her continued support and exaltation, or just putting up with me while I was writing this manuscript. I am especially indebted to those who provided advices, software and other gimmicks that helped the completion of this thesis. I owe Peter Toft as I borrowed some of his L A T E X style, and his si.m matlab visualisation function among other things. L A T E X also got me indebted to Peter S.K. Hansen and Jan Larsen. I am also pleased to thank Xavier Driancourt and Neuristique S.A. for lending me a version of the TLisp/SN software while I was away, and taking care of my endless requests on top of it. This document was written using a customised version of the L A T E X book format with packages fancyhdr, graphicx, apalike and makeidx. Introduction Notation Here are a number of symbols that are used throughout the text, …
منابع مشابه
On-board Component Fault Detection and Isolation Using the Statistical Local Approach
We describe both the key principles and real application examples of a uniied theory which allows us to perform the on-board incipient fault detection and isolation tasks involved in monitoring for condition-based maintenance. We stress that, when designing detection algorithms , the main conceptual task is to select a convenient estimating function. ml, ls, iv and subspace identiication method...
متن کاملDistributed Services in a Messenger Environment: the Case of Distributed Shared-memory I Distributed Services in a Messenger Environment 1
TH ESE pr esent ee a la Facult e des Sciences de l'Universit e de Gen eve pour obtenir le grade de Docteur es sciences, mention informatique par Murhimanya MUHUGUSA de Bukavu (Za re) Th ese No 2903 Gen eve 1997 La Facult e des sciences, sur le pr eavis de Messieurs J. Harms, professeur ordinaire et directeur de th ese (D epartement d'informatique), C. Tschudin, docteur et codirecteur de th ese ...
متن کاملLa programmation loique par contraintes pour l'aide à l'enseignant
Cet article soutient le point de vue que les Environnements Informatis es d'Apprentissage Humain (EIAH) ne consid erent pas as-sez l'enseignant comme un utilisateur potentiel distinct de l'apprenant. Aussi, le n eologisme \pr eceptoriel" est introduit pour d ecrire des EIAH ayant une approche centr ee sur le professeur. Le domaine des logiciels de construction de gures g eom etriques est en-sui...
متن کاملApport des agents mobiles à l 'exécution répartie
Ce papier pr esente une vue d'ensemble de la technologie du code mobile. Il situe ce mod ele d'ex ecution distribu ee par rapport aux mod eles classiques (communication par messages, appel de proc edure distante, objets r epartis, evaluation distante). Il d ecrit l'infrastructure syst eme n ecessaire au support d'agents mobiles pour l'ex ecution et la migration des agents, a la r ealisation de ...
متن کاملStatistical learning for image-based personalization of cardiac models. (Apprentissage statistique pour la personnalisation de modèles cardiaques à partir de données d'imagerie)
Cette thèse porte sur un problème de calibration d’un modèle électromécanique de cœur, personnalisé à partir de données d’imagerie médicale 3D + t ; et sur celui — en amont — de suivi du mouvement cardiaque. Les perspectives à long terme de la simulation personnalisée de la fonction cardiaque incluent l’aide au diagnostic et à la planification de thérapie, ainsi que la prévention des risques ca...
متن کامل