A data mining approach to predict companies’ financial distress

Abstract

Financial distress and companies’ failure have always been a complicated and intriguing problem for businesses. Because of the unfavorable impacts of financial distress on companies and societies, accounting and finance researchers around the world are thinking of ways to anticipate corporate financial distress. Several models are provided in the literature for predicting financial distress. This research develops nonlinear decision tree and linear discriminant analysis models to predict financial distress of companies listed in Iranian Stock Exchange during 2010 to 2015. The drivers are firms’ financial ratios, intellectual capital and performance indicators. According to the results, intellectual capital and financial performance indices have no informational content in decision tree model. Comparing the result show that both models predict financial distress with 90.9% and 81.8% accuracy, respectively. Moreover, the difference between the accuracy of the models however is not meaningful. In other words, two models were very close to each other in terms of predictive power.