RF-PCA2: An improvement on robust fuzzy PCA

Gyeongyong Heo, Kwang Baek Kim, Young Woon Woo, Seong Hoon Kim

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Principal component analysis (PCA) is a well-known method for dimensionality reduction while maintaining most of the variation in data. Although PCA has been applied in many areas successfully, one of its main problems is the sensitivity to noise due to the use of sum-square-error. Several variants of PCA have been proposed to resolve the problem and, among the variants, robust fuzzy PCA (RF-PCA) demonstrated promising results, which uses fuzzy memberships to reduce noise sensitivity. However, there are also problems in RF-PCA and convergence property is one of them. RF-PCA uses two different objective functions to update memberships and principal components, which is the main reason of the lack of convergence property. The difference between two objective functions also slows convergence and deteriorates the solutions of RF-PCA. In this paper, a variant of RF-PCA, called improved robust fuzzy PCA (RF-PCA2), is proposed. RF-PCA2 uses an integrated objective function both for memberships and principal components, which guarantees RF-PCA2 to converge on a local optimum. Furthermore, RF-PCA2 converges faster than RF-PCA and the solutions are more similar to desired ones than those of RF-PCA. Experimental results with artificial data sets also support this.

Original languageEnglish
Title of host publicationIntelligent Information and Database Systems - 4th Asian Conference, ACIIDS 2012, Proceedings
Pages351-360
Number of pages10
EditionPART 2
DOIs
StatePublished - 2012
Event4th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2012 - Kaohsiung, Taiwan, Province of China
Duration: 19 Mar 201221 Mar 2012

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume7197 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference4th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2012
Country/TerritoryTaiwan, Province of China
CityKaohsiung
Period19/03/1221/03/12

Keywords

  • convergence property
  • fuzzy membership
  • noise sensitivity
  • principal component analysis

Fingerprint

Dive into the research topics of 'RF-PCA2: An improvement on robust fuzzy PCA'. Together they form a unique fingerprint.

Cite this