Improved Automatic Clustering Using a Multi-Objective Evolutionary Algorithm with New Validity Measure and Application to Credit Scoring

In data mining, clustering is one of the important issues for separation and classification with groups like unsupervised data. In this paper, an attempt has been made to improve and optimize the application of clustering heuristic methods such as Genetic, PSO algorithm, Artificial bee colony algorithm, Harmony Search algorithm and Differential Evolution on the unlabeled data of an Iranian bank with the credit scoring approach. A survey was also used to measure the clustering validity index which resulted in a new validity index. Finally, the results were compared to identify the best algorithm and validity measure (Das & Konar, 2009).