A Comprehensive Approach for Market Segmentation and Customer Classification Using Data Mining Methods and Alternative Linear Programming

Since market segmentation is essential to develop and implement marketing strategies, has always been focus of marketing researchers’ attention. The advent of new technology and emerging E-businesses, which accumulating immense customer data in their databases, made the market segmentation more fascinating for researchers. Since they found customer database as one of the most valuable asset that if being managed and manipulated effectively, can provide useful knowledge about the customers and prospects. There are numerous non-coherent methods of customer clustering and classification which has been proposed in different disciplines. This paper utilizes different clustering and classification methods and compares their performance in order to propose a comprehensive and integrated algorithm for e-businesses to exploit their databases in a competent manner. In the first step, regency, frequency, monetary (RFM) data is used to ensemble K-Means, Self-Organizing-Map (SOM) and Two-Step clustering methods and Silhouette index is used to evaluate the cluster quality. In second step, different classification methods vis-a-vis multi-group discriminate linear programming (MDLP) are used to compare the performances of the methods in terms of percentage of correct classification. The results show that the performance of MDLP is better than other methods. The problem of insufficient data in databases for classification purpose is also takes into account and fuzzy Delphi method is proposed to select the required data.