The Research On Customer Credit Risk Evaluation Of Nanjing Z Rural Micro Credit Company Based On Neural Network

In the process of rapid development and expansion, Z Micro credit Company increasingly finds it lacks of credit risk management, and urgently needs to build suitable customer credit risk evaluation model, and improve the existing credit risk management level. Currently, Z Micro finance Company is quite subjective in the process of customer credit rating. As a result, its credit rating is improper and eventually it fails to regain the loans on time. Based on the status quo, the thesis took the customer credit risk in Z Micro finance Company as the research object, and established customer credit risk evaluation model by means of BP neural network so as to improve the level of its credit risk management, reduce the possibility of customer default risk,and promote its sustainable development.The thesis involved the following several aspects. Firstly, it started from the basic theory of credit risk evaluation, described the characteristics of Micro finance in China, summarized the existing micro credit risk evaluation index,and formed the theoretical analysis of micro credit risk evaluation. Secondly, the thesis referred international classic micro credit evaluation model,took the firsthand data obtained through field investigation and credit files as the basis,selected the indexed by using coefficient variation analysis and correlation analysis,and established the evaluation indexes of customer characteristic dimension and loan characteristic dimension. At the same time,the thesis researched and determined the key technology of credit risk evaluation that was suitable for the data characteristic of Nanjing Z Micro finance Company,including the initial variables identification,variable assignment and variable dimension reduction techniques,and then established its credit risk evaluation index system. Thirdly,the thesis introduced BP neural network learning algorithm,compared and summarized the characteristics of artificial neural network, and laid theoretical basis for the subsequent empirical research part. Fourthly, the thesis constructed BP neural network model and its algorithm suitable for Z Micro finance Company, completed the run program of BP neural network model through Matlab7.0 software. Besides, it established customer credit risk evaluation model for the Company based on BP neural network, confirmed the effectiveness and applicability of BP neural network for the field of micro finance customer credit risk evaluation and could provide a good basis for Z Micro finance Company to identify customer credit risk. Fifthly, in order to more effectively popularize and apply BP neural network credit risk evaluation model and reduce the micro credit consumer credit risks, the thesis proposed some corresponding use and development suggestions according to the existing problems in the credit risk evaluation of the Company.