Prediksi Kemampuan Pembayaran Klien Home Credit Menggunakan Model Random Forest, Decision Tree, Dan Logistic Regression

Abstract

Home Credit is a global financial company that provides consumer loan services. The purpose of this research is to predict the ability of clients to pay in order to make it easier for companies to provide loans or not. Not being careful in analyzing lending will cause credit risk. So to reduce these risks, the company needs an analysis to predict the client's repayment ability to determine whether to pay or not as a reference for the company in providing credit loans. By using the previous member criteria data, predictions of the smoothness of payments can be made using data mining. The data mining techniques used are Random Forest Classifier, Decision Tree Classifier, and Logistic Regression Classifier. The models used are Random Forest, Decision Tree, Logistic Regression, which determine the likelihood or opportunity based on the data of previous members, and the results. The criteria used consist of the ten best features selected based on the results of best feature importance. The evaluation results of the random forest model are able to predict the ability to pay home credit clients with a high level of test accuracy score of 0.9967, ROC value of 0.9967, recall value of 1.00 compared to the other two models.