Propensity Model Using Decision Trees (LightGBM) for the Management of the Effective Credit Product in a Financial Entity

Roman-Concha U.; López A.; Ruiz-Carrasco K.; Chavez-Herrera C.; Cano D.M.; Piedra J.; Woolcott J.C.; Navarro C. · Advances in Artificial Intelligence and Machine Learning (2025) · DOI: 10.54364/aaiml.2025.51184

Resumen

The objective of this paper was to develop a propensity model based on decision trees (Light-Gbm) for the management of the Credit product in a financial institution. The CRIPS-DM methodology was used as a framework and Python/LightGBM was used for the development of the solution. As a result, it was possible to increase by 5% the effectiveness in credit for each month on a park of 200 thousand customers of the financial institution, which ensures the understanding of the applied model.

Instituciones: Universidad Nacional Mayor de San Marcos; Universidad Tecnológica del Perú; Universidad Nacional de Juliaca