In a volatile economic environment, financial decision making is always associated with risk. Bankruptcy, as one of the most important risks, has a significant impact on the interests of the firm's stakeholders, so presenting appropriate bankruptcy forecasting patterns is of the utmost importance. In this study, after reviewing the theoretical literature and selecting the financial ratios used in previous bankruptcy prediction models as the variable input of the initial model, the financial ratios were adjusted based on the price index and then, using the LARS algorithm, the ratios that have the highest ability to differentiate between bankrupt and non-bankrupt firms were identified, and finally, using the SVM and Naive Bayesian algorithms, the final bankruptcy prediction model was developed. For this purpose, the data of 50 companies listed in Tehran Stock Exchange who had experienced bankruptcy for at least one year from 2008 to 2018 under Article 141 of the Commercial Code were used. The results show that the adjusted financial ratios based on the price index in the model presented by SVM algorithm can be a very good predictor for bankruptcy of companies with an accuracy of 99.4%.
