Identifikasi Kolektibilitas Kredit Menggunakan Decision Tree

Abstract

Credit is a loan agreement to borrow money, in which the bank as a creditor and a debtor as the client. The purpose of this study is to predict the level of collectibility of loans using the web-based decision tree. The variables used for this study include gender, type of loan, the principal amount, purpose, time, code usage, collectability, collateral, monthly installments and monthly pay date. In this research the data used are divided into 3000 training data and 699 testing data. By using J48 decision tree method, the data training was processed to produce the identification model. The model was then tested using 699 testing data. The result showed that 597 data could be identified correctly. Hence the accuracy of the model was 85%.