Session 9 : Recent Developments in the Application of Pairwise Overlap

Pairwise overlap is defined as the sum of misclassification probabilities calculated for a pair of mixture components. It is useful for studying the systematic performance of clustering algorithms. Currently, existing methods allow simulating elliptically distributed data according to pre-specified overlap characteristics. An approach to simulating skewed clusters with a desired overlap is proposed. Next, an extension to measuring overlap in cluster-weighted models is discussed.