Font Size: a A A

Research On Commercial Bank's Customer Classification Based On Data Mining

Posted on:2010-12-23Degree:MasterType:Thesis
Country:ChinaCandidate:J T ZhaoFull Text:PDF
GTID:2189360275974260Subject:Business management
Abstract/Summary:PDF Full Text Request
With the dust and heat of the market,Banks have incurred enormous pressure; therefore,Customer Relation Management (CRM) becomes more and more important in Bank management. Each kind of CRM is based on customer segmentation. Which sorts of customers are beneficial to the bank?Such characteristics include the cost,profit of bank and profit rate of all kinds of service,marketing strategy and service methods.Since it is necessary for the banks in China to make customer segmentation,this paper takes a commercial bank in Chongqing as an example,and chooses data mining method for segment to build a customer segmentation model in order to guide CRM of the bank. The precondition of segmentation model is to analysis the current labeled customers. However,most of the banks in China use a simple way to identify the customers,such as annual salary of the customer or the amount of the saving,and this kind of classification methods are not accurate. Therefore,we adopt a clustering method to cluster the customers and conform the type of the customers before building a segmentation model. Based on that we built a segmentation model via decision tree method,and used it on the future customers,to help classify the customers effectively and reasonably,providing a powerful support tool for CRM in banks.This paper study how to identify important customers of CRM in bank systematicly by data mining technology and statistic technology,import time series element and use SAS to build a customer segmentation model for a commercial bank. In the process of building the model,we adopted SOM clustering method to cluster the customers,and then adopted triple decision tree classifier and binary decision tree combined classifier to build customer classify model,and via evaluating effect among different classifier. At last we got the finial model,and used the model guide the practice.The result of the study had theoretical value and practice value in Customer Relation Management of commercial bank.
Keywords/Search Tags:Customer Relation Management, Data Mining, Self-organizing Mapping, Decision Tree, Combined Classifier
PDF Full Text Request
Related items