TY - GEN
T1 - Another variant of robust fuzzy PCA with initial membership estimation
AU - Heo, Gyeongyong
AU - Kim, Seong Hoon
AU - Woo, Young Woon
AU - Lee, Imgeun
PY - 2011
Y1 - 2011
N2 - Principal component analysis (PCA) is a well-known method for dimensionality reduction and feature extraction. PCA has been applied in many areas successfully, however, one of its problems is noise sensitivity due to the use of sum-square-error. Several variants of PCA have been proposed to resolve the problem and, among the variants, improved robust fuzzy PCA (RF-PCA2) demonstrated promising results. RF-PCA2, however, still can be affected by noise due to equal initial membership values for all data points. The fact that RF-PCA2 is still based on sum-square-error is another reason for noise sensitivity. In this paper, a variant of RF-PCA2 called RF-PCA3 is proposed. The proposed algorithm modifies the objective function of RF-PCA2 to allow some increase of sum-square-error and calculates initial membership values using data distribution. RF-PCA3 outperforms RF-PCA2, which is supported by experimental results.
AB - Principal component analysis (PCA) is a well-known method for dimensionality reduction and feature extraction. PCA has been applied in many areas successfully, however, one of its problems is noise sensitivity due to the use of sum-square-error. Several variants of PCA have been proposed to resolve the problem and, among the variants, improved robust fuzzy PCA (RF-PCA2) demonstrated promising results. RF-PCA2, however, still can be affected by noise due to equal initial membership values for all data points. The fact that RF-PCA2 is still based on sum-square-error is another reason for noise sensitivity. In this paper, a variant of RF-PCA2 called RF-PCA3 is proposed. The proposed algorithm modifies the objective function of RF-PCA2 to allow some increase of sum-square-error and calculates initial membership values using data distribution. RF-PCA3 outperforms RF-PCA2, which is supported by experimental results.
KW - KD-tree
KW - Membership initialization
KW - Nearest neighbor
KW - Noise sensitivity
KW - Principal component analysis
UR - http://www.scopus.com/inward/record.url?scp=84872103130&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-20042-7_13
DO - 10.1007/978-3-642-20042-7_13
M3 - Conference contribution
AN - SCOPUS:84872103130
SN - 9783642200410
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 120
EP - 129
BT - Intelligent Information and Database Systems - Third International Conference, ACIIDS 2011, Proceedings
T2 - 3rd International Conference on Intelligent Information and Database Systems, ACIIDS 2011
Y2 - 20 April 2011 through 22 April 2011
ER -