Presenting A Model for Predicting the Financial Crisis of The Iranian Capital Market Using Hybrid Algorithms

Managers and investors always tend to predict the results of their decisions and investments based on expectations and existing conditions. Undoubtedly, for this purpose, they need a correct analysis of the current situation and predict future events. Therefore, the aim of the current research is to provide a dynamic model for predicting possible financial crises. In order to achieve the purpose of the research, first, by using the content analysis, 25 indicators were identified from the categories of macroeconomic indicators, industry factors, company characteristics, political, cultural, and behavioral events. Then, by using multivariable regression algorithm, smart ant colony algorithms and particle swarm optimization, and using the combined data of 173 companies accepted in Tehran Stock Exchange from 2010 to 2020, the proposed model was tested. The findings based on the regression method showed that some internal and external criteria had a significant impact on the financial crisis of companies. the findings showed that in terms of efficiency, the optimization method of ants is the most efficient in the problem of predicting the financial crisis. Finally, it was found that the information of independent variables can predict the financial crisis of companies. The findings of the research also show that up to five years before the financial crisis, it is possible to predict the financial crisis in companies with relatively high accuracy, but as the financial crisis subsides, due to the decrease in the clarity and accuracy of the financial sector forecasting indicators, the predictive ability of the model also decreases.