Preface
This research was completed in May 2004 and submitted to University of Pennsylvania as partial fulfillment of requirements for a PhD degree in Operations and Information Management. The first part of the research focuses on using behavior patterns for customer segmentation. It advances data mining theory by presenting a novel pattern-based clustering approach for customer segmentation. The second part of the research studies free-shipping promotions on the Internet. Since May 2004, there have been academic contributions and developments in industry that are especially relevant to this research.
First, Xiong et al. (2005) points out that there has been considerable interest in using association patterns for clustering. Although several interesting algorithms have been developed, further investigation is needed to characterize the benefits of using association patterns and the most effective way of using them for clustering. They present a new clustering technique – bisecting K-means Clustering with pAttern Preservation (K-CAP) – which exploits key properties of the hyperclique association pattern and bisecting k-means. Experimental results on