| Summary: | 碩士 === 國立臺北科技大學 === 經營管理系碩士班 === 100 === In this thesis, the information about the wealth management customers of a bank was analyzed by making use of RFM (Recency, Frequency, Monetary) and profitability analysis, and active customers were clustered by applying Two Step cluster method in IBM SPSS Modeler. First, the RFM values of individual customers were derived from transaction data and M value (purchase amount) was replaced by customers’ profitability. Then the RFM values, composite RFM score, average purchase amount, percentage of periodic purchase, and age were assigned as input variables for SPSS’s Two Step clustering. The results show that there were five groups (i.e. clusters) of customers with different behavior patterns being successfully identified. Besides illustrating the transaction behaviors of each group, this research compared the dissimilarity among clusters and suggested cluster names. The proposed clustering procedure can be used to correctly evaluate and analyze customer transaction features and contributions, as well as for empirical analysis and regular review on profitability of customers.
|
|---|
