Article (Scientific journals)
Cellwise Robust regularized discriminant analysis
Aerts, Stéphanie; Wilms, Ines
2017In Statistical Analysis and Data Mining, 10, p. 436–447
Peer reviewed
 

Files


Full Text
JGLDA2.pdf
Author preprint (328.83 kB)
Request a copy

All documents in ORBi are protected by a user license.

Send to



Details



Keywords :
Cellwise robust precision matrix; Classification; Discriminant analysis; Penalized estimation
Abstract :
[en] Quadratic and Linear Discriminant Analysis (QDA/LDA) are the most often applied classification rules under normality. In QDA, a separate covariance matrix is estimated for each group. If there are more variables than observations in the groups, the usual estimates are singular and cannot be used anymore. Assuming homoscedasticity, as in LDA, reduces the number of parameters to estimate. This rather strong assumption is however rarely verified in practice. Regularized discriminant techniques that are computable in high-dimension and cover the path between the two extremes QDA and LDA have been proposed in the literature. However, these procedures rely on sample covariance matrices. As such, they become inappropriate in presence of cellwise outliers, a type of outliers that is very likely to occur in high-dimensional datasets. In this paper, we propose cellwise robust counterparts of these regularized discriminant techniques by inserting cellwise robust covariance matrices. Our methodology results in a family of discriminant methods that (i) are robust against outlying cells, (ii) cover the gap between LDA and QDA and (iii) are computable in high-dimension. The good performance of the new methods is illustrated through simulated and real data examples. As a by-product, visual tools are provided for the detection of outliers.
Disciplines :
Mathematics
Author, co-author :
Aerts, Stéphanie ;  Université de Liège > HEC Liège : UER > UER Opérations : Informatique de gestion
Wilms, Ines;  Katholieke Universiteit Leuven - KUL > Faculty of Economics and Business > Leuven Statistics Research Centre (LStat)
Language :
English
Title :
Cellwise Robust regularized discriminant analysis
Publication date :
2017
Journal title :
Statistical Analysis and Data Mining
ISSN :
1932-1864
eISSN :
1932-1872
Publisher :
Wiley, Hoboken, United States - New Jersey
Volume :
10
Pages :
436–447
Peer reviewed :
Peer reviewed
Name of the research project :
contract number 12M8217N
Funders :
FWO - Fonds Wetenschappelijk Onderzoek Vlaanderen [BE]
Available on ORBi :
since 23 December 2016

Statistics


Number of views
102 (22 by ULiège)
Number of downloads
5 (5 by ULiège)

Scopus citations®
 
4
Scopus citations®
without self-citations
3
OpenCitations
 
3

Bibliography


Similar publications



Contact ORBi