Identification of the most critical factors in bankruptcy prediction and credit classification of companies

Abstract

Banks and financial institutions strive to develop and improve their credit risk evaluation methods to reduce financial loss resulting from borrowers’ financial default. Although in previous studies, a lot of variables exploited from financial statements had been used as the input to the bankruptcy prediction process such as financial ratios, seldom a machine learning method base on computing intelligence used to selection the most critical of them. In this research, the data from companies which were listed in Tehran`s stock exchange and OTC market during 26 years since 1992 to 2017 have been investigated as population and 218 companies have been selected as sample, and the method of an ant colony optimization algorithm with k-nearest neighbor have been used to feature selection and classify the companies. In this study, the problem of imbalanced dataset has been solved with sampling technic. The results have shown that variables such as EBIT to total sales, equity ratio, current ratio, cash ratio and debt ratio are the most effective factors in predicting the health status of companies. The accuracy of final research model is estimated the bankruptcy prediction ranges between 75.5% to 78.7% for the training and testing sample.