A Random Forest Approach to Appraise Personal Credit Risk of Internet Loans

In view of the fact that in recent years, internet loan business has gradually exposed that the pre prevention and management of risks are not comprehensive enough, which has led to the untimely response of most platforms to the consequences of the borrower's breach of contract, resulting in insufficient cash flow on the platform, resulting in a series of problems such as cash withdrawal difficulties and serious runs. In this study, the borrower's personal credit risk identification is studied, and the data mining process and method of credit data risk are proposed. Select the Internet loan data of a domestic city commercial bank, and use random forest algorithm and decision tree algorithm to identify and predict the risk. The research results show that the prediction accuracy of the random forest model built in this paper reaches 97% through the high-quality pre-processing of the original credit data, indicating that the model has a high reliability and can well identify the risks related to Internet loans of commercial banks. At the same time, the research also finds that several interesting characteristics, such as the borrower's balance, amount and fund use, are crucial to identify whether the borrower defaults. In general, the research in this paper can improve the level of commercial banks' lending decision-making, and contribute to the sound development of commercial banks.