Investigation And Forecasting Of Tehran Stock Exchange Indices According To Macroeconomic Variables: Using Artificial Neural Network With Multilayer Perceptron (Mlp)

Abstract

Background: Due to the key impact of macroeconomic variables on stock market indices, their nonlinear and nonparametric behavior, investors, financial managers and economic actors are in macro risk conditions. Therefore, predicting the volatility of indicators has always been one of the most controversial issues in the field of financial issues and is very important.

Method: In order to predict the fluctuation of these indices in Tehran Stock Exchange and the impact of macroeconomic variables such as exchange and inflation rates, gold and oil prices, export and import volumes, relevant statistical data were extracted during 2014-2018. Then, stock indices were forecasted during the mentioned period. In this study, artificial neural network with multilayer perceptron (MLP) was used.

Results: The results show that macroeconomic variables affect the Tehran Stock Exchange index. Neural networks have the ability to predict the stock market index in time periods with an appropriate error rate.