Mining important association rules on different customer potential value segments for life insurance database

To maximize customer profitability, companies should exert effort to acquire new customers, as well as to retain existing customers and add value. An efficient way of achieving such goals is to explore and profile customers' past purchase behavior and mine out their possible further needs and wants. When faced with increasingly diversified consumption demands, customers should be segmented based on their potential and willingness to purchase. Therefore, in this study, we propose the customer potential value (CPV) matrix for the segmentation of applicants based on the degree of their potential value and their willingness to buy for an insurance database. In this proposed CPV matrix, the applicants will be categorized into four dimensions defined as the opened group, desire-deficiency group, perception-deficiency group, and closed group. Furthermore, we use the data mining technique to determine the important association rules for each segment of the CPV matrix. The results show that more powerful support of association rules can be obtained via the segmentation of customers based on the CPV matrix.