Application Of Machine Learning Combinatorial Model In Personal Credit Assessment

In today's society,personal consumer credit business has developed rapidly,and the choice of effective methods to control the risks in the business process has attracted much attention in the industry.Therefore,credit assessment is an effective tool for banks and financial institutions to guide loan decisions correctly and obtain profits.In the field of machine learning,ensemble learning method is a method developed in the field of credit assessment in recent years,which plays an indispensable role in the improvement of model performance,and has advantages for accurately identifying good borrowers and bad borrowers.This paper follows the steps of ensemble learning,put forward a kind of credit assessment model: XGboost-Logistic combinatorial model,which is based on the stacking algorithm.In this paper,stacking algorithm is divided into two layers of learners: primary learners and secondary learners.Primary learners is the combinatorial classifier of Logistic regression and XGBoost,secondary learners is Logistic regression classifier.Taking the consumer credit data on Kaggle platform as an example by using Python programming.The data is firstly preprocessed,and then feature engineering is carried out to screen out the necessary features.Then,a single Logistic regression model,XGboost classification model,XGboost-Logistic combinatorial model based on bagging algorithm and a combinatorial model based on stacking algorithm are established respectively.The results show that,compared with the other three models,The XGBoost-Logistic combinatorial model based on Stacking algorithm takes into account the characteristics of high accuracy,strong robustness and better interpretation,and has certain reference significance in real life.Although the performance of the selected optimal model is only a little better than that of other models,in real life,for financial institutions and other similar institutions,even a little improvement in efficiency can bring huge benefits to the institution.Therefore,this model is of great practical significance in the real life business decisions of organizations.