Font Size: a A A

The Research And Application Of Data Mining Technology For The Bank Customer Relationship Management

Posted on:2009-05-08Degree:MasterType:Thesis
Country:ChinaCandidate:Y ZhangFull Text:PDF
GTID:2178360272957907Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Customer relationship management in banking, which centers on customer, is gradually drawing everyone's attention. The grate customer relationship management is capable of bringing about tremendous revenue. Data mining can anticipate customer behavior, and support people decision-making accordingly. How to make use of data mining technology to effectively provide decision-making information for managers is the key of customer relationship management system for commercial banks.From the point of view of CRM application and development, this paper introduces data mining technology, and brings forward a blue print to control credit risk forecast of CRM in banks, and makes a prediction for customer credit risk. First of all, it starts with customer relationship management and data mining; introduces the definition and structure of customer relationship management, the main methods in data mining technology and its application in customer relationship management. Current application for data mining technology and its application to customer relationship management in banking are analyzed; expatiates upon the internal relationship between data mining and customer relationship management in banking. In the aspect of selecting data mining methods, this paper investigates Decision Trees Algorithm specially. The algorithm has successfully applied to many classification issues. Try to adopt its ID3 algorithm to implement credit risk forecast for customer data in banking. According to the simulated basic customer data in banking, a decision tree model is built to forecast credit risk for customers and finds out classification rules.. In the process of relation, explains the methods and procedures of pretreatment for tested data in detail,database module,generate decision-tree module,validate and forecast module. In the meanwhile, because ID3 algorithm requires discrete attributes, the algorithm creates a series of splits for continues attributes, and pre-treats these data to satisfy with requirement of algorithm. As some recognition issues may occur during algorithm implementation, it puts forward an attribute identification method. Furthermore, combines with the characteristics of commercial banks, it builds up a credits risk forecast model for banks, validates the validity of model, and generates the corresponding rules from the model. This model can be applied to the control of customer credits risk for customer relationship management in banking. At last, it makes full use of conclusions that comes from rules and combines with the basic information of customer to implement customer credits risk prediction.
Keywords/Search Tags:data mining, customer relationship management, decision tree, credit risk
PDF Full Text Request
Related items