PCA-Guided k-Means with Variable Weighting and Its Application to Document Clustering

Authors:
Katsuhiro Honda;Akira Notsu;Hidetomo Ichihashi
Affiliations:
Osaka prefecture University, Osaka, Japan 599-8531;Osaka prefecture University, Osaka, Japan 599-8531;Osaka prefecture University, Osaka, Japan 599-8531
Venue:
MDAI '09 Proceedings of the 6th International Conference on Modeling Decisions for Artificial Intelligence
Year:
2009

Citing 8
Cited 0

Principal component analysis based on a subset of variables: variable selection and sensitivity analysis

MSI '00 Proceedings of the conference, volume III (in honor of Professor Minoru Siotani on his 70th birthday on Multivariate statistical analysis
Pattern Recognition with Fuzzy Objective Function Algorithms

Pattern Recognition with Fuzzy Objective Function Algorithms
Linearized cluster assignment via spectral ordering

ICML '04 Proceedings of the twenty-first international conference on Machine learning
K-means clustering via principal component analysis

ICML '04 Proceedings of the twenty-first international conference on Machine learning
Automated Variable Weighting in k-Means Type Clustering

IEEE Transactions on Pattern Analysis and Machine Intelligence
Linear fuzzy clustering techniques with missing values and their application to local principal component analysis

IEEE Transactions on Fuzzy Systems
Regularized Linear Fuzzy Clustering and Probabilistic PCA Mixture Models

IEEE Transactions on Fuzzy Systems
Linear Fuzzy Clustering With Selection of Variables Using Graded Possibilistic Approach

IEEE Transactions on Fuzzy Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

PCA-guided k -Means is a deterministic approach to k -Means clustering, in which cluster indicators are derived in a PCA-guided manner. This paper proposes a new approach to k -Means with variable selection by introducing variable weighting mechanism into PCA-guided k -Means. The relative responsibility of variables is estimated in a similar way with FCM clustering while the membership indicator is derived from a PCA-guided manner, in which the principal component scores are calculated by considering the responsibility weights of variables. So, the variables that have meaningful information for capturing cluster structures are emphasized in calculation of membership indicators. Numerical experiments including an application to document clustering demonstrate the characteristics of the proposed method.