BP Neural Network’s Application In Rural Credit Cooperatives Customer Classification

Rural credit cooperatives are also an integral part of finance, but with thedevelopment of China’s market economy, the key to rural credit cooperatives as themain users are experiencing unprecedented financial risks. World Bank studies showthat credit risk is the main cause of bank failures reasons.Therefore, this articlemainly through data mining technology, based on customer credit risk, customers areclassified for credit unions, credit unions so as to provide information on the client’smanagement decision-making basis. In this paper, based on data mining of customercredit classification algorithm mainly includes the following several aspects.Firstly, based on BP neural network for credit unions customer classificationmodel to study, and for the existence of the BP neural network is easy to fall intolocal optimal solution, convergence is slow and difficult to configure the networkstructure and other shortcomings, combined with genetic algorithm, we proposed theconcept of genetic neural networks and genetic neural network model for the designof the study. Secondly, combined with a credit Hunan customer data, combined withthe credit union’s own characteristics, customers on credit risk factors were analyzed,based on the analysis of customer data by means of genetic neural network model willcredit the customer rating as good, fair, qualified, unqualified, worsening five levels,to achieve the customer’s credit classification. Through examples prove, studied inthis paper based on genetic neural network classification model credit unioncustomers with better customer segmentation results.In this paper, scholars using neural network model based on customersegmentation studies, use of neural networks for simple customer segmentationalgorithm when deficiencies, with the advantage of genetic algorithm, for customerclassification model improvements. And tests have shown by example, studied in thispaper based on genetic neural algorithm credit customer classification model has arelatively better customer classification accuracy, able to provide better customermanagement credit methodological support.