Using the Imperialistic Competitive Algorithm Model in Bankruptcy Prediction and Comparison with Genetic Algorithm Model in Listed Companies of Tehran Stock Exchange

Abstract

Bankruptcy prediction is a major issue in classification of companies. Since bankruptcy is extremely costly, investors, owners, managers, creditors, and government agencies are interested in evaluating the financial status of companies. This study tried to predict bankruptcy among companies registered in Tehran Stock Exchange (Iran) by designing imperialist competitive algorithm and genetic algorithm models. It then compared the accuracy of the two models in financial conditions of Iran and sought the best model to predict company bankruptcy one, two, and three years before its incidence. Also uses a model to surveying the financial position and also the subject of continuing operations about them to improve the quality of decision taken by shareholders and stakeholders. The study sample consisted of 38 bankrupt and 38 non-bankrupts companies during 2007-2016. The final variables used in both algorithms were five financial ratios. The results showed that the imperialist competitive algorithm had better accuracy than the genetic algorithm in bankruptcy prediction at the mentioned intervals.