Social Credit Evaluation And Regulation Research Based On Agent

Coupled with economic development, the increase in credit fraud has become the dominant contributor to the criminal endeavors. Nowadays, group-buying and shopping online is rather prevailing among people. Meanwhile, the number of credit complaints has increased dramatically. In addition, under the impact of international financial crisis and European sovereign debt crisis, the credit default problems frequently occur in various sectors. This credit problems not only disorder the normal international market and delay the progress of market-oriented reform, but also improve the transaction cost in the market. Particularly, the government is establishing the social credit system, and have promulgated a lot of relevant laws and regulations. Meanwhile, many scholars also study and explore the related issues of the social credit, while this issues are most qualitative analysis. Based on the existing credit research literatures, this paper complete a comprehensive literature review, and discuss the social credit supervision and credit risk assessment.Firstly, this paper provides a comprehensive review of the social credit literature from the perspectives of theoretical foundation, scoring methods, and regulatory mechanisms. The study found that (1) artificial intelligence models have higher accuracy than traditional model; (2) supervision mechanism study stay in the qualitative analysis.Secondly, this paper proposed a credit risk assessment model based on aritificial intelligence Agent (ELM-C). Different from existing models using a fixed cutoff value (0.0 or 0.5), the proposed classification model especially considers the optimal cutoff value as one important evaluation parameter in credit risk modeling, to enhance the assessment accuracy. From the results, main conclusions can be summarized as follows:(1) Comparing the five benchmarks without cutoff selection, the single ELM model outperforms all considered AI tools, confirming the superiority of the ELM model in terms of prediction accuracy. (2) The novel ELM-C method with cutoff selection performs the best amongst all listed models, indicating the effectiveness of the proposed model. The results further indicate that the novel method can be used as one promising classification tool for credit risk assessment, with high prediction accuracy. (3) Selecting the optimal cutoff value is significant to capture the data features, especially for asymmetric data set. The optimal cutoff can improve the classification accuracy of credit risk assessment modle.Thirdly, to explore the impact of government regulation on firms'profit and defective rate, this paper formulated a multi-agent-based model which include three economic agents:government, firms and customers. The simulation results indicate the following:(1) Government regulation is necessary to firms'developments. Appropriate government regulation can help enterprises develop rapidly, while excessive government regulation will otherwise inhibit firm's development. (2) The information disclosure regulation policy is more effective to increase profit and decrease defective rate than penalty regulation policy. (3) Firm'profit is sensitive to information disclosure ratio. A slight variation in the information disclosure ratio has a significant influence to the enterprise credit rating.This paper provides a comprehensive review of the social credit literature from the perspectives of theoretical foundation, scoring methods, and regulatory mechanisms. Furthermore, in terms of credit risk assessment, the proposed model based on artificial intelligence Agent is superior to other model. In terms of credit regulatory mechanism, credit simulation model based on Multi-Agent shows that the government regulation is good for the long-term development of the enterprise, and information disclosure regulation policy is more effective than penalty policy.