Poster (Scientific congresses and symposiums)
Robust discriminant analysis based on the joint graphical lasso estimator
Aerts, Stéphanie; Croux, Christophe; Wilms, Ines
2016Leuven Statistics Days
 

Files


Full Text
Poster_AERTSStephanie.pdf
Publisher postprint (79.59 kB)
Request a copy

All documents in ORBi are protected by a user license.

Send to



Details



Keywords :
Joint graphical lasso; Discriminant analysis; Cellwise contamination
Abstract :
[en] Linear and Quadratic Discriminant Analysis (LDA/QDA) are the most often applied classification rules under the normality assumption. When there is not enough data, the quadratic rule, which requires the estimation of one precision matrix in each class, is often replaced by the linear one, based on the homoscedasticity assumption. This strong assumption is however rarely verified in practice and ignores the intrinsic différences between groups that may be of particular interest in the classification context. In this aper, alternatives to the usual maximum likelihood estimates for the precision matrices are proposed that borrow strength across classes while allowing for heterogeneity at the same time. This results in a classifier that is intermediate between QDA and LDA. Moreover, our estimator is sparse: the undesirable effect of uninformative variables is reduced. The performance of the method is illustrated through simulated and real dataset examples.
Disciplines :
Mathematics
Author, co-author :
Aerts, Stéphanie ;  Université de Liège > HEC-Ecole de gestion : UER > UER Opérations : Informatique de gestion
Croux, Christophe;  Katholieke Universiteit Leuven - KUL > Faculty of Economics and Business > ORSTAT
Wilms, Ines;  Katholieke Universiteit Leuven - KUL > Faculty of Economics and Business > ORSTAT
Language :
English
Title :
Robust discriminant analysis based on the joint graphical lasso estimator
Publication date :
October 2016
Number of pages :
A0
Event name :
Leuven Statistics Days
Event organizer :
Leuven Statistics Research Centre, KUL
Event place :
Leuven, Belgium
Event date :
du 20 au 21 octobre 2016
Funders :
KU Leuven - Katholieke Universiteit Leuven [BE]
Available on ORBi :
since 18 October 2016

Statistics


Number of views
76 (3 by ULiège)
Number of downloads
2 (1 by ULiège)

Bibliography


Similar publications



Contact ORBi