DocumentCode :
457381
Title :
Class Separability in Spaces Reduced By Feature Selection
Author :
Pranckeviciene, Erinija ; Ho, TinKam ; Somorjai, Ray
Author_Institution :
Inst. for Biodiagnostics, Nat. Res. Council Canada
Volume :
3
fYear :
0
fDate :
0-0 0
Firstpage :
254
Lastpage :
257
Abstract :
We investigated the geometrical complexity of several high-dimensional, small sample classification problems and its changes due to two popular feature selection procedures, forward feature selection (FFS) and linear programming support vector machine (LPSVM). We found that both procedures are able to transform the problems to spaces of very low dimensionality where class separability is improved over that in the original space. The study shows that geometrical complexities have good potentials for comparing different feature selection methods in aspects relevant to classification accuracy, yet independent of particular classifier choices
Keywords :
computational complexity; computational geometry; feature extraction; linear programming; pattern classification; support vector machines; class separability; classification accuracy; forward feature selection; geometrical complexity; linear programming support vector machine; Biomedical measurements; Councils; Geometry; Humans; Interference; Labeling; Linear programming; Space technology; Support vector machine classification; Support vector machines;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Pattern Recognition, 2006. ICPR 2006. 18th International Conference on
Conference_Location :
Hong Kong
ISSN :
1051-4651
Print_ISBN :
0-7695-2521-0
Type :
conf
DOI :
10.1109/ICPR.2006.365
Filename :
1699514
Link To Document :
بازگشت