Predicting Tehran Stock Exchange Index Based On The Effect Of Macroeconomic Variables On The Stock Index And Using Artificial Neural Network With Multilayer Perception

Abstract

Given the key role of macroeconomic variables and their impaction the Stock Exchange index, and
considering the nonlinear behaviour and nonparametric of the Stock Exchange index, investors,
financial managers and economic agents will practically experience high risks, so predicting the index
trend as one of the most controversial issues in the financial affairs is of crucial importance. In order
to analysis the Exchange index, the statistical data of the Tehran Stock Exchange, the macroeconomic
variables of foreign exchange, gold, inflation, oil, export, import were extracted from 2014to2018, and
then the prediction of the stock exchange index was conducted over these years. Artificial neural
Network with a multilayer perception (MLP) structure was used in this study. The results indicate that:
the macroeconomic variables affect Tehran Stock Exchange index; and neural networks have the ability
to predict a stock index in time intervals with acceptable error rate