Use of machine learning techniques in the prediction of credit recovery

Abstract

This paper is an extended version of the paper originally presented at the International Conference on Machine Learning and Applications (ICMLA 2016), which proposes the construction of classifiers, based on the application of machine learning techniques, to identify defaulting clients with credit recovery potential. The study was carried out in 3 segments of a Bank's operations and achieved excellent results. Generalized linear modeling algorithms (GLM), distributed random forest algorithms (DRF), deep learning (DL) and gradient expansion algorithms (GBM) implemented on the H2O.ai platform were used.