Analysis of Distinct Feature Groups in the Credit Scoring Problem

TL;DR: In this article, the importance of new feature groups not commonly employed for the credit scoring task and others already used was analyzed, such as historical, geolocation, web behavior, and demographic data.

Abstract: Registration and financial data have been traditionally used for the credit scoring problem. However,slight improvements in the reliability of the scores positively impacts financial companies. Therefore, exploring newfeatures is a strategic task. This work analyzes the importance of new feature groups not commonly employed forthe credit scoring task and others already used. We categorized features from open credit scoring datasets, suchas German and Australian and compared their groups with the ones of a company dataset used in this work. Ourdataset contains unusual feature groups, such as historical, geolocation, web behavior, and demographic data. In ouranalyzes, we first conducted bivariate tests with each feature-pair to assess their individual importance. Secondly, weran XGBoost machine learning model with each feature group to evaluate each group importance. We also appliedfeature selection with binary Particle Swarm Optimization to assess the groups importance when combined. Next, weemployed correlation tests to find inner and inter-correlation among the features groups. Finally, we used the companydataset and employed AdaBoost, Multilayer Perceptron, and XGBoost algorithms to find the best model for the task.Some of our main findings were that the unusual features added a slight improvement to registration features. We alsodetected reasonable inner correlation among some feature groups and found that all groups were relevant for the taskwith the Historical Group as the most promising. Lastly, XGBoost obtained the best performance over AdaBoost andMultilayer-perceptron for the task.