Research On Credit Rating Of Chinese Life Insurance Corporation Based On BP-neural Network Model

Honesty as the foundation is the precondition for the healthy progress of the lifeinsurance companies. It is also an important factor to enhance the competitiveness of lifeinsurance companies and establish a good social image for life insurance industry. However,Information asymmetry existed between the consumer and the insurance company, resultingin low consumer confidence in life insurance companies, life insurance companies sufferedthe credit crisis. Credit ratings generated from asymmetric information theory. Credit rating isthe intrinsic quality of the enterprise’s comprehensive testing and evaluation, enabling the lifeinsurance company’s solvency comprehensive evaluation; it is highly professional andtargeted credit analysis. Credit rating for the life insurance companies can provide consumerswith accurate information about the risks, enabling consumers to grasp the real situation oflife insurance companies, increasing consumer’ trust for life insurance companies, andpromoting the development of the life insurance industry.In this paper firstly much domestic and foreign relevant literature is reviewed andsummarized, and the basic theory related to this research is further analyzed. It provides thetheory support for later research. Chinese life insurance company’s credit rating is analyzedtheoretically on the credit rating theory, asymmetric information theory, a comprehensive riskmanagement thinking and credit factor. It provides the theory support for later research.Secondly in reference to the basis of the foreign life insurance company’s Credit rating indexsystem,in the light of the distinguishing feature life insurance companies in China, the lifeinsurance company’s credit rating index system in China is established, and analyzes itsapplicability. It laid the foundation for further research. Once again, the traditional BP neuralnetwork model is analyzed simply, and then the paper discusses the improved BP neuralnetwork model developed on the basis of the tradition. In this paper with the improved BPneural network model, according to the various indexes of fifteen life insurance company andactual rating result, by Matlab training and simulation, A network model can be obtained. Itsoutputs are very similar to the actual results. The overall error between the network modeloutput and the actual rating results can satisfy the demands so that the life insurancecompany’s credit rating model is trained. Finally, through an example it verified accuracy and applicability of the trained network model for life insurance companies credit rating. theresults of training model show that the trained network rating results fit the company’s actualsituation. It accurately reflects the company’s credit rating. Therefore it can be predicted bythe model of life insurance company credit rating.