PRINCIPAL COMPONENT ANALYSIS BASED FEATURE SELECTION FOR CLUSTERING

摘要：

Feature Extraction (FE) methods have been proved to be very effective for dimension reduction, but the features attained are meaningless. In order to exploit the effectiveness of FE methods to support Feature Selection (FS), this paper proposed a new FS approach for clustering based on Principal Component Analysis (PCA) called PS. It first uses PCA to transform the data from original feature space into a new feature space whose features are linear combination of the original ones, and then evaluates the importance of the original features based on the newly generated features and the feature importance measure proposed in this paper, finally selects features incrementally according to their importance to improve the performance of the clustering algorithm. Experiment is carried out on several popular data sets and the results show the advantages of the proposed approach.

关键词： Feature selection Principal component analysis Clustering

作者: JUN-LING XU BAO-WEN XU WEI-FENG ZHANG ZI-FENG CUI

作者单位: School of Computer Science and Engineering, Southeast University, Nanjing 211189, China School of Computer Science and Engineering, Southeast University, Nanjing 211189, China State Key La Department of Computer, Nanjing University of Posts & Telecommunications, Nanjing 210003, China

会议类型: 国际会议

会议名称: 2008 International Conference on Machine Learning and Cybernetics(2008机器学习与控制论国际会议)

会议地点: 昆明

会议语种:英文

页码: 460-465

在线出版日期: 2008-07-12（万方平台首次上网日期，不代表论文的发表时间）