Abstract

The main research purpose of this article is to sort out consumer information without specific personal data and use the target company’s neural network risk control decision model which is based on the risk scoring trajectory. We find the main interpretation variables, and then verify the direction and extent of the target’s influence to verify empirical results and help the target company make a correct decision analysis on the credit applicants’ loans. The final empirical results show that the analysis results of the minimum square method and the analysis results of the truncated regression model can be found: the gender, population identity characteristics, consumer performance capabilities, consumer risk of trust, consumer preferences, consumer behavior characteristics, behavioral characteristics, number of financial institutions the consumer have applied for loans, credit litigation, and loan period were the nine main explanatory variables affecting the composite score. Regarding the influence of each variable on the comprehensive score: the impact of four index variables such as identity characteristics, contract performance ability, consumption preference, and behavioral characteristics on the comprehensive score is positive; it means that with higher identity feature score, contract performance ability, consumer preference score, and behavioral characteristic score, the comprehensive score increases.