| Supply chain finance(SCF)has effectively mitigated the financing problems of small and medium-sized enterprises(SMEs)since its birth,becoming one of the major businesses of banks.Nowadays,online supply chain finance(OSCF)has become an emerging mode of SCF along with the popularization of Internet technology.Therefore,it is urgent to find an effective way to evaluate credit risks of OSCF.Due to differences in risk factors and business models in different industries,this essay takes auto industry,whose SCF is the most mature,as the research objects to build a more targeted indicator system and conduct evaluation for the credit risks of OSCF.This essay firstly clarifies the concepts,characteristics and models of traditional SCF and OSCF,introduces classical theories of credit risk,analyzes the sources of credit risks in different modes of OSCF,and compares various credit risk evaluation models.Secondly,according to the summary of indicators in the previous literature,this essay builds an innovative indicator system for OSCF credit risk evaluation from the perspectives of macro environment,financing enterprise and core enterprise qualification,etc.Thirdly,in the empirical part,the Lasso-logistic model,the multi-layer perceptron(MLP)model and the combined model of both are adopted.Through comparing the prediction results of three models,the combined model proves to maintain high accuracy while perform better in robustness than MLP model.Thus the combined model is more suitable for evaluating the credit risk of OSCF of auto industry.In the end,this essay comes to conclusions and attempts to suggest the government,banks and financing companies on controlling credit risks of OSCF. |