THE FINANCIAL DISTRESS PREDICTION MODELLING USING COMBINATION OF LOGISTIC REGRESSION AND DECISION TREES

Abstract

This article examines the prediction of financial distress among small and medium-sized enterprises (SMEs). The primary research question is whether integrating logistic regression with decision tree techniques enhances the accuracy and practical relevance of bankruptcy prediction compared to logistic regression alone. To address this question, both a standard logistic regression model and a hybrid model combining logistic regression with decision tree–based variable selection were applied to a dataset of 2,770 accounting records from 986 Czech companies, of which 238 were declared insolvent between 2010 and 2016. The results indicate that while logistic regression independently provides valuable insights, its hybrid integration improves model robustness and predictive performance. The combined approach not only enhances the accuracy of bankruptcy prediction but also identifies key financial distress indicators and their critical threshold values. These findings suggest that the hybrid method can serve as a useful analytical tool for financial analysts, policymakers, and business managers seeking reliable early-warning mechanisms for financial distress in SMEs.