Research And Application Of Small And Medium Sized Enterprises Credit Risk Assessment System

Medium-sized and small enterprises play an important role in promoting the growth of the national economy. In recent years, the state has also issued a series of policies to encourage and guide the banking institutions to credit support for SMEs, to ensure the smooth development of SMEs. However, the bank loans to SMEs also face the risk of uncollectible loans. How to assess and control credit risk the bank has become an urgent problem to be solved.On the basis of in-depth study of risk assessment, rough sets and support vector machines and other related theories, in view of the problems in the process of credit risk assessment of small and medium enterprises, such as too many indexes, small sample size, nonlinear classification and so on and combined with the actual enterprise credit risk assessment system, the SME credit risk assessment model based on rough set and support vector machine is proposed to simplify the evaluation index system of enterprise credit risk assessment and to improve the accuracy of risk assessment. In terms of credit risk evaluation index screening, firstly to collect more than 300 small and medium enterprises sample data from a commercial bank, according to the expert advice and experience knowledge preliminary from these many enterprises the operational properties of the selected liquidity ratio, quick ratio, cash ratio, asset liability ratio, interest coverage ratio, ratio of net assets, return on assets and other 20 enterprises credit risk evaluation index. Preprocessing the 20 indexes data of enterprise sample, that is to do data fill and continuous attribute discretization operation for the enterprise credit risk assessment information sheet to construct the risk assessment decision table. Then, according to the evaluation decision table, the approximate accuracy attribute reduction algorithm in rough set theory is used to reduce the 20 evaluation indexes, under the premise of ensuring the effective information of the decision table is not lost, the flow rate, the asset liability ratio, the net assets ratio and the cost and expense ratio are obtained, which greatly simplify the evaluation system. On the construction of enterprise credit risk assessment model, Firstly, based on the evaluation index, the paper constructs the feature vector of the enterprise to evaluate the sample. After the enterprise characteristic sample data is normalized, the linear kernel function, S kernel function and other kernel functions are calculated and compared, through their effect on the risk assessment of enterprise feature samples, Gauss kernel function is used as the kernel function of enterprise risk assessment model. The penalty parameters and kernel parameters of the model were determined by the grid search method and the multi fold cross validation method. Finally, an enterprise risk assessment model based on support vector machine is built, which solve the credit evaluation system in the presence of the assessment indicators are more, the sample size is small, nonlinear classification and other issues and improve the accuracy and effectiveness of the credit risk assessment.This paper through to a prefecture level city small and medium-sized enterprise carries on the risk appraisal application, verifying the validity of the credit risk assessment system of small and medium enterprises constructed in the paper and realizing the digital enterprise credit risk assessment, to make the assessment more effective and scientific.