| In recent years,with the implementation of national innovation driven strategy and introduction of a number of policies and measures to promote the rapid development of China's innovative small and micro businesses,innovative small and micro businesses is becoming an important channel of entrepreneurship to enrich people and important foundation for national economic and social development.It plays a very important role in expanding the employment of the local public,increasing people's income and tax,improving the level of people's lifeand improving the vitality of the market economy.However,the financing has been an important bottleneck to hinder the development of innovativesmall and micro businesses.In order to better solve the financing problem of small and micro enterprises,commercial banks must establish a complete set with a credit rating mode in accordance with the actual situation of innovative small and micro enterprises.This paper regards the actual situation of A bank “loan linkage business” in Hunan province as the research background.It selects a number of innovative small and micro businesses from credit system of A bank in Hunan Province.According to the credit system it selects rating index of enterprises,and obtains rating data through the Internet.Aiming at the defects of the existing credit rating system,it analyses effectively how to improve loans accessibility of innovative small business and to reduce the operation cost of the commercial bank.It also puts forward the new solution of credit rating model,and builds two new model of credit rating respectively using BP neural network algorithm and AdaBoost combination classification method.In addition,through the score resultsof professional credit personnel for the sample enterprise and combination with the advice of the professional credit personnel in A bank of Hunan Province,the sample enterprise credit level is divided into three categories I,II and III,and the credit level threshold is determined.The highest levelI illustrates that A bank in Hunan province can invest the enterprise in'equity +debt mode;II level shows that A bank in Hunan province cancarry out credit business on the enterprises;III level says that the enterprise credit rating is so low that the A bank of Hunan Province in such enterprises will not carry out its credit business without collateralprovided by the enterprise or guarantees provided by the third party.From the empirical evaluation results,the two types of credit rating model can better solve the problem of credit rating of innovative small and micro enterprises in Hunan A bank,and provide a feasible solution for it.Finally,the paper analyses on the contrast of the rating results of the two credit rating models and ones of professional personnel in A bank for the sample enterprise.From the results of the two models we can see thatalthough the rating efficiency of BP neuralnetwork model in the certain degree is higher than that of AdaBoost combination classification model,AdaBoost the combination classification model is significantly better than the BP neural network model in three aspects of accuracy,stability and risk control.Therefore,this paper believes that credit rating model ofAdaBoost combination classification algorithm is more practical for the credit business of small and micro businesses credit business in A bank of Hunan Province,and it has strong promotion effect on classification performance.It has certain guiding function for commercial bank credit personnel to carry out credit analysis in the practical application,and can provide good support for the credit decision of A bank in HunanProvince. |
PDF Full Text Request