TY - GEN
T1 - RF-PCA2
T2 - 4th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2012
AU - Heo, Gyeongyong
AU - Kim, Kwang Baek
AU - Woo, Young Woon
AU - Kim, Seong Hoon
PY - 2012
Y1 - 2012
N2 - 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.
AB - 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.
KW - convergence property
KW - fuzzy membership
KW - noise sensitivity
KW - principal component analysis
UR - http://www.scopus.com/inward/record.url?scp=84863347341&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-28490-8_37
DO - 10.1007/978-3-642-28490-8_37
M3 - Conference contribution
AN - SCOPUS:84863347341
SN - 9783642284892
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 351
EP - 360
BT - Intelligent Information and Database Systems - 4th Asian Conference, ACIIDS 2012, Proceedings
Y2 - 19 March 2012 through 21 March 2012
ER -