Research And Realization Of The SMEST Risk Assessment Model Based On Industry Risk Coefficient

Risk is everywhere, the rapid economic development of China has stimulated the activity of financial markets, risk management of financial activities not only becomes an important topic in the field of financial science, economics and other related fields, but also in computer science and machine learning. Due to the Small and Mid-sized Enterprises of Science and Technology(SMEST) become a highlight branch to drive economic transformation and upgrading through scientific and technological innovation. However, the development of SMEST receive financing difficult restrictions, intrinsic factor is the lack of credit records,and the external reason, open channel is hard to get a set of perfect and targeted designed evaluation system to assess the risk of SMEST, which limits the accuracy of the credit evaluation. So many financial institutions will be preference for large enterprises, which form in capital raising work even more of a trouble of SMEST. If we want to solve the financing problems of SMEST, we must solve its risk assessment, but the risk assessment method of traditional Small and Medium Enterprises is not suitable for SMEST. To solve the problem,this paper tailoring a credit evaluation index system for SMEST, then a targeted credit rating model is built on the basis of the index system. This unique targeted can ensure the accuracy of the assessment, therefore this topic research has practical significance.For the establishment of the credit rating index and related modeling method, the writer read a large number of literature. On the basis of the understanding of domestic and foreign research situation, this article summarized and analyzed the advantages and disadvantages about the existing credit evaluation and related technology, and make a further research against insufficient points, the research content mainly involves the following aspects:(1) First, pointed out that different industries have different influence for model analysis.Use the ISM model on the influencing factors of SMEST risk assessment, and find main influence factors in addition to the financing risk is risk industry, so the industry difference can’t be ignored.(2) Then, Optimized index system. In view of the existing credit evaluation index system of SMES is too general, lack of industry pertinence. Optimized existing index system by introducing the new index "industry risk coefficient". This new index includes two parts, theindustry’s own risk coefficient and the risk in phase of the life cycle of the industry.(3) Once more, Construct risk assessment model based on the optimized index system.The model applied decision tree for risk assessment, and through the introduction of the global mutual information + redundancy, namely, the mutual information between the properties and classification category, redundancy of properties at the same time. The traditional decision tree algorithm in the choice of property was improved by choosing mutual information, which overcome the shortcoming of local optimum of the traditional decision tree using information gain,and improve the accuracy.(4) Last, through the empirical analysis to verify the feasibility and effectiveness of the improvement of decision tree of risk assessment model based on the new index system.