Issue No. 03 - March (2009 vol. 21)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/TKDE.2008.163
Tianyi Jiang , Stern School of Business/New York University, New York
Alexander Tuzhilin , New York University, New York
On the Web, where the search costs are low and the competition is just a mouse click away, it is crucial to segment the customers intelligently in order to offer more personalized products and services to them. Traditionally, customer segmentation is achieved using statistics-based methods that compute a set of statistics from the customer data and group customers into segments by applying distance-based clustering algorithms in the space of these statistics. In this paper, we present a direct grouping based approach to computing customer segments that groups customers in terms of optimally combining transactional data of several customers to build a predictive model of customer behavior for each group. We consider customer segmentation as a combinatorial optimization problem of finding the best partitioning of the customer base into disjoint groups and show that finding an optimal customer partition is NP-hard. We propose several suboptimal direct grouping segmentation methods, empirically compares them against traditional statistics-based hierarchical and affinity propagation based segmentation, and 1-to-1 methods across multiple experimental conditions. We show that the best direct grouping method builds mostly small sized customer segments and significantly dominates the statistics-based and 1-to-1 approaches across most of the experimental conditions, while still being computationally tractable.
Personalization, Clustering, classification, and association rules, Data mining, Clustering
T. Jiang and A. Tuzhilin, "Improving Personalization Solutions through Optimal Segmentation of Customer Bases," in IEEE Transactions on Knowledge & Data Engineering, vol. 21, no. , pp. 305-320, 2008.