Title: Building a robust linear model with forward selection and stepwise procedures
Authors: Khan, Jafar A ×
Van Aelst, Stefan
Zamar, Ruben H #
Issue Date: 2007
Series Title: COMPUTATIONAL STATISTICS & DATA ANALYSIS vol:52 issue:1 pages:239-248
Abstract: Classical step-by-step algorithms, such as forward selectio
n (FS) and stepwise (SW)
methods, are computationally suitable, but yield poor resu
lts when the data contain
outliers and other contaminations. Robust model selection
procedures, on the other
hand, are not computationally efficient or scalable to large d
imensions, because they
require the fitting of a large number of submodels. Robust and
efficient versions of FS and SW are proposed. Since FS and SW can
be expressed in
terms of sample correlations, simple robustifications are o
btained by replacing these
correlations by their robust counterparts. A pairwise appr
oach is used to construct
the robust correlation matrix – not only because of its compu
tational advantages
over the
-dimensional approach, but also because the pairwise approa
ch is more
consistent with the idea of step-by-step algorithms. The prop
osed robust methods
have much better performance compared to standard FS and SW.
Also, they are
computationally very suitable and scalable to large high-di
mensional datasets.
ISSN: 0167-9473
Publication status: published
KU Leuven publication type: IT
Appears in Collections:Non-KU Leuven Association publications
× corresponding author
# (joint) last author

Files in This Item:
File Description Status SizeFormat
RobFSandSW_report.pdf Published 389KbAdobe PDFView/Open


All items in Lirias are protected by copyright, with all rights reserved.

© Web of science