The Research And Application About The Customer Classification And Sale Oriented To CRM System In The Bank

With the developing of economic, the competition between the banks isincreasingly fierce. The traditional strategy of commercial banks has been graduallyreplaced by a customer-centric, service target strategy. Bank majorly use customerrelationship management (CRM) to manage customers and services. The data israpidly expanding in geometric scale in this information generation. Vast amounts ofcustomer data have been taking into account of the commercial banks. The bank’scustomers are very different. It is necessary to classify customers and researchrelated product, so that the bank can accurately grasp the status of existingcustomers to recommend a variety of products. At the same time on differentcategories of customers, bank takes different marketing policies.For the important role of customer classification and marketing, thecombination of theoretical and operation is applied in this paper.First introduce the use of the CRM system in the commercial banks and theapplication of CRM in the bank and summarize the CRM theory and the data miningtheory. Establish the data mining procedure model for the PCRM mining system.Second establish the topics of customer classification and association degreebetween products and customers. Preprocess the data in the Bank A’s PCRM systemsand build customer classification model and associated model with C5.0, C&RT andApriori algorithm on the platform of Clementine.And then use Cognos to display the estimation results at the front-end.At last evaluate the performance for the both classic decision tree algorithmC5.0and C&RT. From the view of entropy and Gini, validate the classic decision treeC5.0and C&RT algorithm. Give suggestions about sales and activities for differenttypes of customers, from product-related marketing view. And give sale process tovaluable customers from the view of customer loss.