Premium
On the Applicability of Multidimensional Classification Procedures
Author(s) -
Werneckk K.D.,
Bäumler H.,
Kalb G.
Publication year - 1983
Publication title -
biometrical journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.108
H-Index - 63
eISSN - 1521-4036
pISSN - 0323-3847
DOI - 10.1002/bimj.19830250604
Subject(s) - statistics , multivariate analysis of variance , feature selection , variance (accounting) , selection (genetic algorithm) , pattern recognition (psychology) , estimation , computer science , mathematics , class (philosophy) , artificial intelligence , accounting , management , economics , business
An example from the EEG‐analysis shows that particularly in using the R‐error estimation the variance analysis and discrimination might indicate contradictory results. Therefore the use of results of MANOVA for a subsequent discrimination appears not to be advisable as a rule. If the number of test persons per class does not amount to at least the five‐fold of the number of characters, the error estimation according to the R‐method is not applicable. Since the feature selection by the methods investigated in this paper only pretend an improvement of the classification, producing even misleading results in single cases, it should not be used, at least not with smaller random samples.