Investigating the effect of macroeconomic variables on the Tehran Stock Exchange index: Comparison of neural network and regression VAR models

Document Type : Original Article

Authors

1 PhD Student, Department of Management, North Tehran Branch, Islamic Azad University, Tehran, Iran.

2 Assistant Professor, Department of Finance, Shahroud University of Technology, Shahroud, Iran.

3 Associate Professor, Department of Mathematics, South Tehran Branch, Islamic Azad University, Tehran, Iran.

Abstract

Many studies in financial science have focused on accurate forecasting with respect to investment risk. The stock market is an institution that collects savings and liquidity from the private sector to finance long-term investment projects. The indicators of this market are influenced by several factors, one of the most important of which is economic variables. Considering the key role of macroeconomic variables and its effect on the stock exchange index, due to the nonlinear and non-parametric behavior of the stock exchange index, investors, financial managers and economic actors will be placed in macro risk conditions. It is one of the most controversial issues in finance, it is very important. The present study compares neural network models and time series on the effect of macro variables on the Tehran Stock Exchange index. Therefore, the multilayer and regression Prostron neural network models of the VAR model have been investigated. Tehran Stock Exchange Index has been selected as a statistical population in the period from the beginning of April 2013 to the end of March 2017. In order to have a criterion for comparing the four criteria of root error, mean square error, mean absolute value of error percentage, average absolute value of error and coefficient of determination have been used. Examination of the designed neural network performance and regression of prediction error criteria showed that the neural network model is superior to the VAR series model in terms of error criteria.

Keywords

Political Sociology of Iran
Volume 5, Issue 9
November 2022
Pages 1472-1489

Files

Share

How to cite

Statistics