Research On Personal Credit Evaluation Based On Multi-Grained Cascade Forest Model

Personal loan business is a huge credit resource of commercial banks,which provides a lot of income for commercial banks,and also plays a role in optimizing the capital structure of banks.Personal credit assessment is the key step for commercial banks to control personal credit risk,which plays an important role in risk identification and risk prediction.However,many commercial banks do not evaluate personal credit in a standard way,and the relevant system is not perfect.It is common to evaluate personal credit customers and potential customers only by business experience.This paper refers to the research results of personal credit evaluation in recent years,and believes that the credit evaluation system needs to analyze the personal financial data of borrowers.In the era of big data,the traditional classification algorithm has been hard to analyze personal financial data,so it is necessary to introduce a new classification algorithm to provide technical support for individual credit evaluation.The credit evaluation system is actually a process of classifying the borrower's credit based on the analysis of the borrower's characteristics.This process includes a series of work,such as collecting and processing data,algorithm selection,feature engineering,model establishment,model evaluation and improvement.Random forest and neural network are machine learning methods commonly used in these classification problems.Neural network,especially the deep neural network model,has obvious advantages in the analysis of high-dimensional nonlinear discrete data,but also has the defects of high complexity and low interpretation.In this paper,we use Multi-Grained Cascade Forest algorithm,which is efficient and has the ability of deep mining data.We study the Multi-Grained Cascade Forest model with Random Forest and Extra-Trees as the basic learners.The results show that the credit evaluation model based on Multi-Grained Cascade Forest model has high accuracy and stability,and is feasible and practical in personal credit evaluation.