USP Electronic Research Repository

A new perspective to null linear discriminant analysis method and its fast implementation using random matrix multiplication with scatter matrices

Sharma, Alokanand and Paliwal, K.K. (2012) A new perspective to null linear discriminant analysis method and its fast implementation using random matrix multiplication with scatter matrices. Pattern Recognition, 45 (6). pp. 2205-2213. ISSN 0031-3203

[img] PDF - Published Version
Restricted to Repository staff only

Download (525kB)

Abstract

Null linear discriminant analysis (LDA) method is a popular dimensionality reduction method for solving small sample size problem. The implementation of null LDA method is, however, computationally very expensive. In this paper, we theoretically derive the null LDA method from a different perspective and present a computationally efficient implementation of this method. Eigen value decomposition (EVD) of SþT SB (where SB is the between-class scatter matrix and SþT is the pseudo in- verse of the total scatter matrix ST) is shown here to be a sufficient condition for the null LDA method. As EVD of SþT SB is computationally expensive, we show that the utilization of random matrix together with SþT SB is also a sufficient condition for null LDA method. This condition is used here to derive a computationally fast implementation of the null LDA method. We show that the computational complexity of the proposed implementation is significantly lower than the other implementations of the null LDA method reported in the literature. This result is also confirmed by conducting classification experiments on several data sets.

Item Type: Journal Article
Subjects: T Technology > TA Engineering (General). Civil engineering (General)
Divisions: Faculty of Science, Technology and Environment (FSTE) > School of Engineering and Physics
Depositing User: Ms Shalni Sanjana
Date Deposited: 25 Jun 2011 04:05
Last Modified: 18 Jan 2017 03:30
URI: http://repository.usp.ac.fj/id/eprint/4814
UNSPECIFIED

Actions (login required)

View Item View Item

Document Downloads

More statistics for this item...