An Ensemble Method for Validation of Cluster Analysis
Clustering is more subjective work than classification and regression. Though classification and regression have many general validation measures, clustering has few validation measures. Also, it is difficult to develop general measure of cluster validation. So, many evaluation measures have been published for cluster validation. In this paper, we propose an ensemble method of validation for cluster analysis. We use voting approach to some validation measures of cluster analysis. To verify our improved performance, we make experiments by some objective data sets from UCI machine learning repository.
Keywords: Cluster Analysis, Cluster Validation, Ensemble Method, Voting, Internal measures, Stability measures
Download Full-Text
ABOUT THE AUTHOR
Sunghae Jun
Cheongju, Chungbuk 360-764, Korea
Sunghae Jun
Cheongju, Chungbuk 360-764, Korea