Font Size: a A A

Churn Warning Analysis On Commercial Bank’s Customer

Posted on:2015-04-12Degree:MasterType:Thesis
Country:ChinaCandidate:P W SunFull Text:PDF
GTID:2309330422471723Subject:Applied statistics
Abstract/Summary:PDF Full Text Request
As business competition intensifies, companies have to increase various means ofcompetition in order to attract customers, and various personalized services have beenoffered for a variety of customer groups. Because customers are facing diversity ofchoice, the corporate loyalty reduces. Customer churn has become a major problem forbusinesses. Whether a company can solve this problem is related to survivability of acompany in such a competitive environment and the company’s prospects. This article isbased on this subsistent issue. Taking the loss of customers to a commercial bank as astudy sample, we use SAS data mining software, statistical knowledge and thecombination of business experience in a variety of senior operational staff to analyzeand predict customer churn problem. We try to analyze the loss of customers’ behaviorcharacteristics with a variety of customer transaction data and to establish theappropriate warning churn prediction model based on various behavioral characteristicsof customers, so that the business people can take the appropriate marketing tools togain time and retain customers for the loss of customers.In this paper, we take a systematic study of the market churn which is concerned tosolve the problem, according to market rules, statistical knowledge and data miningtechnology. Firstly, we examine and analyze the standard process to solve this problem,and study the key and difficult problems for each process. Secondly, we put forward aset of solutions to solve the problem with the actual data of enterprises and the types ofproblems we are facing. Thirdly, based on various data mining processes and carefuldata preparation, we select the variables associated with the target in many variables.Finally, we use principal component analysis and logistic regression to solve someproblems existed in these variables. In addition, we compare the model results, and testthe real effect of the whole modeling process. We identify the main features of customerchurn, and fight for valuable time for customer retention.
Keywords/Search Tags:Data mining, Models of customer churn, Principal component analysis, Logistic regression, Test of significance
PDF Full Text Request
Related items