Evaluating the Effect of Content of Inflation Accounting Information in Comparison with Historical Information in Designing Bankruptcy Prediction Models Based on Traditional and Meta-Innovative Approaches

Bankruptcy Prediction is one of the branches of finance that has received more attention in in recent research as bankruptcy patterns have been developed. In most of the researches in the field of Prediction the financial performance of companies and in particular, bankruptcy, only Predicting or comparing the predictive power of models using historical information of financial statements has been done. Since historical accounting information has been used more in Iran, the main purpose of this study is to consider the effects of inflation on input variables in designing a bankruptcy prediction model. Therefore, the variables in design of two different models were classified into two groups of financial ratios, adjusted and historical. Then, the ratios were identified using the LARS algorithm that had the highest ability to differentiate between bankrupt and non-bankrupt companies. Finally, the final bankruptcy prediction model was designed using the logit regression test and SVM and Naive Bayesian algorithms. For this purpose, the data of 50 companies listed on the Tehran Stock Exchange were used, which had experienced bankruptcy according to Article 141 of the Commercial Code. The results of this study indicate that the financial ratios adjusted based on the price index are more suitable predictor for corporate bankruptcy. Also, the bankruptcy prediction model designed by SVM algorithm can be a very good predictor for corporate bankruptcy with 99.4% accuracy.