Bankruptcy prediction using machine learning models: empirical results in the Colombian manufacturing industry (2018-2022)

Abstract

The purpose of this study is to examine financial indicators that reveal the situation of corporate failure in manufacturing companies in Colombia. The use of these indicators is based on previous studies that have used predictive models of corporate fragility: multiple discriminant analysis, logistic regression, and machine learning. This work uses logistic regression and random forests models. This work is based on financial indicators made of the data reported between 2018 and 2022 in the database Sistema de Información y Riesgos Empresariales (SIREM) of the Superintendence of Companies.