Customer Credit Rating by Machine Learning

Abstract

Recently, people's consumption attitudes have also changed, being inclined to spend in advance. Banks and other financial institutions use credit rating models as a tool to evaluate the credit score of individuals, determine whether to grant the loan to the applicant. One of the biggest challenges for the banking industry in assessing the customers’ credit is that it is unlikely to provide a manual review to classify them because of the huge volume of data on applicants. Therefore, it is necessary to establish a suitable and effective credit rating model to help banks evaluate the quality of applicants. This paper focuses on the problems existing in the development of personal credit rating system and tries to find the best solution in the field of personal credit rating system. By selecting independent variables that are highly correlated with delinquency behavior, using different models for testing, and comparing the results of the models, this paper finally draws the conclusion that different algorithms combined by the group decision method can make better decisions.