Study On Credit Evaluation Of Small Enterprises In Manufacturing Industry Based On Non-parametric Bayes Discrimination

With the rapid development of China's economy and society,small enterprises,as an important part of China's enterprise group,have become an indispensable solid force to promote national economic growth;and manufacturing industry,as the leading industry in China's economic development,directly reflects the level of national productivity.Small manufacturing enterprises not only promote China's economic development,but also provide a lot of work.Job placement has alleviated the difficulty of employment in China.However,due to the imperfect financial system,imperfect management mechanism and the lack of credit evaluation system for small manufacturing enterprises in commercial banks,it is difficult for small manufacturing enterprises to finance.In order to solve this problem,it is particularly important to establish a scientific and complete credit evaluation system for small manufacturing enterprises.This paper introduces the development status,financing status and credit evaluation status of small manufacturing enterprises in China,combs the research results at home and abroad,and finds that scholars mostly study the credit evaluation under parameters.Considering that most of the index data in reality do not obey normal distribution,non-parametric method is chosen to study.In this paper,165 manufacturing lending enterprises of a commercial bank are selected as research samples.Forty-four indicators are selected to form the primary index system,which takes account of both financial and non-financial indicators.The first round of screening is carried out by non-parametric Bayes discriminant method,and the indicators that can significantly identify the default status of enterprises are selected.The second round of screening is carried out by non-parametric clustering analysis method,avoiding the residual index.Standard response information redundancy,constructed a complete and non-repetitive index system consisting of 16 indicators,such as asset-liability ratio and operating profit margin.The final credit index system not only corresponds to the international "5C" principle,but also improves the discriminant accuracy from 73.1%to 80.6%.According to the influence degree of each index on the discriminant accuracy in the non-parametric Bayes discriminant results,the index is empowered.After empowernent,the proportion of financial indicators is 0.45,and that of non-financial indicators is 0.55,which proves that non-financial indicators are important to manufacturing industry.Credit evaluation of small enterprises plays an important role.Finally,the credit evaluation model is obtained by weights and standardized index data,and compared with the credit evaluation model constructed by logistic regression.The results show that the credit evaluation model constructed by non-parametric Bayes discriminant method is better,and the accuracy of discrimination reaches 81.6%.This study provides a reference basis for commercial banks to evaluate the credit of small manufacturing enterprises,and gives policy recommendations in terms of industry standards,laws and regulations.It also reduces the financial risk of banks and promotes the financing of small manufacturing enterprises.