Study On Practicing Data Ming In Customers' Relation Management In Bank

MIT's scientific and technological review chose the ten new-developing technologies which would exert a great influence to mankind. Among them the data excavation occupied the third. Data excavation method was initially used in bank and telecommunication industries and the data preservation of the former is more intact than that of the latter. This thesis attempts to use the judgment-tree in data excavation method to set up judgment-tree model in the aim of finding out the distinctive characteristics of housing-loan applicants among banks' customers and then it advocates different management ways in accordance of different customers' characteristics. At the same time this thesis puts forward an improved judgment-tree which will be used to help banks practice new data excavation tasks for customers' relations management.Judgment-tree method has been applied to solve lots of categorized problems successfully but so far in China it has never been applied into classification and prediction about customers' behavior of banks. This thesis tries to adopt judgment-tree algorithm to classify and predict customers' data of banks. The purpose lies in the exploration of the method's technological problem in operating banks' data. And through the comparative analysis with the method of neural network, judgment-tree method excavates the potential characteristics of banks' customers and therefore it guides the management of customers' relation of banks.This thesis is divided into four chapters:Chapter one: management theory of customers' relation. It mainly introducesthe theory of customers' relation management, the concept of data excavation and the interrelationship between them. It also raises the study subject of this thesis.Chapter two: Given the data available and this data excavation task, it makes a study method comparison between neural network and judgment-tree. It confirms the latter as the model and presents a brief statement on judgment-tree.Chapter three: According to the basic principles of judgment-tree in data excavation, combing the concrete data of bank branches in middle China, it probes into setting up judgment-tree data model on the foundation of data pretreatment and concept division layer so as to predict customers in group and to test the model aswell.Chapter four: It explains and evaluates the established judgment-tree model. Based on judgment result it points out how to give adoptable suggestions for banks to their different customers in adherence of this model. Meanwhile it proposes an improved judgment-tree method.