Sum-of-Squares Based Cluster Validity Index and Significance Analysis

Different clustering algorithms achieve different results to certain data sets because most clustering algorithms are sensitive to the input parameters and the structure of data sets. Cluster validity, as the way of evaluating the result of the clustering algorithms, is one of the problems in cluster analysis. In this paper, we build up a framework for cluster validity process, meanwhile a sum-of-squares based index for cluster validity purpose is proposed. For homogeneous data based on independent variables, the proposed clustering validity index is effective in comparison to some other commonly used indexes. We use resampling method in the framework to analyze the stability of clustering algorithm, and the certainty of cluster validity index also.