Study On The Application Of Clustering And Discriminant Analysis Method In Bank Customer Classification

This paper puts forward the topic "Study on the Application of clustering and discriminant analysis methods in bank customer classification" for the problems faced by commercial bank account managers in maintaining customers,such as many customers,difficult management and large workload.The methods of clustering analysis and discriminant classification are introduced in detail,and the effectiveness of the method is verified.The first chapter introduces the background and significance of the research,and expounds the status quo of "big data","cluster analysis","discriminant analysis" and "commercial customer management system of commercial banks" at home and abroad.This chapter also predicts the prospects of the topic and illustrates the main research content of this paper.The second chapter introduces the clustering analysis algorithm in detail,and introduces the detailed process of clustering analysis from the two aspects of hierarchical clustering and direct clustering,and further explains it by using the matrix model in higher mathematics.A simple clustering algorithm model was also established and implemented using matlab.Then the clustering experiment of bank individual customers is used to verify the effectiveness of clustering analysis.The third chapter is mainly about the description of the discriminant analysis algorithm.The discriminating process of distance discriminant method,Bayesian discriminant method and Fisher discriminant method is introduced in detail.At the same time,the corresponding algorithm models are established for these three discriminant methods,and they are implemented by matlab.Then,through the discriminant experiment on the bank's individual customers,the correctness of the discriminant analysis is verified.The fourth chapter combines cluster analysis algorithm and discriminant analysis algorithm to introduce examples for verification.Through the processing of randomly collected bank personal customer financial asset data,the idea of clustering existing customers and then discriminating new customers was carried out.The experimental results confirm that the account manager can use the cluster analysis and discriminant classification methods to judge the feasibility of the customer's financial preference,and also verify the validity of the classification and maintenance of the customer.Finally,the conclusion that the clustering and discriminant analysis method is applicable to the classification of commercial bank customers is obtained.It is foreseen that cluster analysis and discriminant analysis techniques will play a more important role in the future production and operation activities of commercial banks.