Measure Of Credit Risk Based On Neural Network Expert System

With the unprecedented competitions in the international financial field and large-scale ofloan combining, old credit measurements can not meet the need nowadays. According tothe situation in China, the reason for banks'low benefits and bad debts is not only thewrong system but also the ignorance of credit risk analyses. Many of the past researchesonly focused on quantitative analysis, which made the using scope so narrow. This papertries to build a credit risk scoring model which integrates quantitative analysis andqualitative analysis together in order to extend the using scope. The paper bases on theknowledge of credit risk, expert system and neural network. Firstly, it briefly introducesthe cause of credit risk and the modern scoring models. Secondly, it determines thenetwork structure, the parameter, the function, method and gist of the division, so as toestablish a credit risk scoring model combined with quantitative and qualitative guide line.Finally, after testing the model, the resuR indicates that this model has an obviouslysuperiority in measuring credit risk.