Customer Credit Rating Evaluation Research Based On Multiple-attribute Decision Making Support Vector Machine

Abstract:Credit evaluation of customer refers to that assessment agencies do quantitative and qualitative analysis for the credit status, solvency and prospects of customers during certain operating period, using scientific, standardized, unified evaluation method. Since the credit assessment of customers directly affect the products and credit limit that commercial banks can provide to them, and the guarantee banks can accept etc., and also determines the credit risk of the loan to some extent, the accuracy and predictability of the credit assessment is particularly important. In conclusion, adapting a scientific and reasonable method for the credit evaluation of customers helps to bring the credit control forward. Banks may be able to anticipate the possible risks of their customers, ensuring the safety of their funds to the maximum extent.After reviewing the history, this passage summarizes the principles and requirements of the credit assessment for the loan provided by the banks. Outline the credit risk, credit assessment and index system from some aspect. And also analyze the main factors which influence the credit status of the corporate. Explain the principles and specific practices for establishing index system of the credit assessment. Introduce the manifestation and index of the credit assessment from the aspect of the establishment of index system for the enterprise credit evaluation. Describe the composition of the index and the parameters from qualitative and quantitative aspect, using expert scoring method to get parameters for qualitative index. Then cope with the data making use of the normalization method and factor analysis. Establish the credit evaluation system model using multi-attribute decision-SVM method, and then describe the specific algorithm and steps. Compared with the conventional distinguishing method, the multi-attribute decision-SVM method using in this paper to establish evaluation model for credit customer is more scientific, objective and universal.In short, the credit evaluation model stabling in this paper is meaningful. The empirical results show that this method is objective, rational and scientific. This method has a positive reference value for banks to assess the risk of their credit customers.4figures,9tables,65references.