A Study of Customer Segmentation Validity Improvement Using Nested-Clustering Method

• Main Authors: Chih-Wei Hsu, 許智為
• Other Authors: 羅淑娟
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/t5gpcm

碩士 === 國立臺北科技大學 === 工業工程與管理系所 === 94 === Nowadays, Since Customization has been used extensively, market needs to realize the most important demand from customer, Customer Segmentation plays an important role in the marketing region. A large amount of research applied clustering analysis to segment the huge customer data in order to provide marketing strategy principle for manager or marketing staffs in recent years. However, the ability of clustering method will affect the result of customer segmentation and directly influence the drawing-up of marketing strategy further.  In this research, we adopt RFM model for segment variables on clustering analysis by using an on-line community web site. Due to the Traditional Clustering Method(K-Means and Fuzzy C-Means) being unable to use all of the variables uniformly result in the final cluster forming a sheet or belt shape distribution, we propose a novel clustering method which segments large numbers of data in the first step, then assigns cluster centroid for the second step executing the final clustering method. By this method, we expect to improve the defective in traditional clustering method and promote the validity of customer segmentation.