The Designing Early Warning System of Financial Crisis Outbreak in Tehran Stock Exchange by Decision Tree

The main purpose of this paper is to predict financial crisis in stock exchange market along with designing warning syetem by data analysis and then to present to financial policy makers for preventing the outbreak or decreasing the effects of crisis. Due to this purpose, it’s used the weekly datas during the years from 10.03.1997 to 03.22.2019. The mean of crises in this present paper is the falling more than 15% of stock price rather to last three months. Hence, it’s been used the dummy variable for operating the dependent variable. It’s been used the residual of autoregressive integrated moving average (ARIMA) for measuring the shocks caused by the index of stock price, exchange rate, price of oil and gold. Based on the results by the different data showed the financial crisis outbreak is the most important variable for predicting the crisis of Tehran stock exchange in weekly data in the last periods. Hence it’s claimed that falling the stock index is affected the value of index in the last periods more than external shock including shock of exchange rate as well as gold and oil. As it’s determined the detection accuracy of crisis is 81.82%, the same for all trees. It means that 36 crisises have been predicted and recognized from the 44 ocurred crisises during the mentioned periods (1121 weeks).