Customer Segmentation Using Overlapping Cluster Algorithm
CRM, a recent developing marketing paradigm, pursues long-term relationship with profitable customers. Most of the firms are now adopting data mining as a strategy way of capturing knowledge about potential needs of targeted customers, future trends in the market and other management factors. Therefore, it is essential to segmenting customers to find out the targeted customers and profitable customers.Clustering is one of the most popular techniques in data mining, which classify n objects into k disjoint clusters. The goal of clustering is to identify distinct groups in a dataset. In this paper, a new overlapping cluster algorithm (OCA) is defined, and customer segmentation using OCA is discussed. The differences between OCA and traditional clustering algorithms are in two respects. First, OCA is overlapping, because clusters are allowed to overlap with one another. Second, the clustering is non-exhaustive, because an object is permitted to belong to no cluster. Because of the variety of customers and customers value assessing, which are overlapping by nature, OCA is useful in customer segmentation.
Data mining Overlapping cluster algorithm Customer Relationship Management
Feng Qian
the Institute of Management Science & Information engineering, Hangzhou Dianzi University, Hangzhou, Zhejiang, 310018, China
国际会议
武汉
英文
2007-09-21（万方平台首次上网日期，不代表论文的发表时间）