Premium
A Cautionary Note on Selection of Variables in Discriminant Analysis
Author(s) -
Murray Gordon D.
Publication year - 1977
Publication title -
journal of the royal statistical society: series c (applied statistics)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.205
H-Index - 72
eISSN - 1467-9876
pISSN - 0035-9254
DOI - 10.2307/2346964
Subject(s) - selection (genetic algorithm) , linear discriminant analysis , statistics , mathematics , feature selection , computer science , artificial intelligence
Summary When searching for a good subset of variables for use in discriminant analysis, one finds that the apparent error rate does not decrease monotonically with increasing size of subset. This paradox is resolved in terms of bias associated with searching through large numbers of subsets. The explanation throws some doubt onto many established techniques of variable selection.