Building Customers` Credit Scoring Models with Combination of Feature Selection and Decision Tree Algorithms

Abstract

Today`s financial transactions have been increased through banks and financial institutions. Therefore, credit scoring is a critical task to forecast the customers’ credit. We have created 9 different models for the credit scoring by combining three methods of feature selection and three decision tree algorithms. The models are implemented on three datasets and then the accuracy of the models is compared. The two datasets are chosen from the UCI (Australian dataset, German dataset) and a given dataset is considered a Car Leasing Company in Iran. Results show that using feature selection methods with decision tree algorithms (hybrid models) make more accurate models than models without feature selection.