INTERNATIONAL JOURNAL OF SOCIAL HUMANITIES SCIENCES RESEARCH (JSHSR)

MODELLING AND PREDICTING STOCK RETURNS IN ISTANBUL STOCK EXCHANGE (ISE): AN ARTIFICIAL NEURAL NETWORK APPROACH
(MODELLING AND PREDICTING STOCK RETURNS IN ISTANBUL STOCK EXCHANGE (ISE): AN ARTIFICIAL NEURAL NETWORK APPROACH )

Author : Fatih MANGIR  ; Zeynep ÖZTÜRK KARAÇOR & Abdul-Razak Bawa YUSSIF  
Type :
Printing Year : 2019
Number : 48
Page : 4470-4477
    


Summary


Keywords


Abstract

Prediction models and techniques are of great ability and importance to all economic agents. More specifically, successful prediction of stock exchange rates and returns are beneficial for investors who want to make higher returns as well as governments in decision-making. Therefore, this study examines the predictive capability of a backward propagation artificial neural network. Moreover, it is aimed at determining whether a backward propagation neural network is capable of accurately predicting the Borsa Istanbul (BIST 100 index) stock returns. Interest rates, inflation rates, exchange rate, money supply, industrial production index of the Turkish economy as well as five world stock market indices are used as input variables to predict returns. The experimental results suggest that the model successfully predicts the monthly return of the BIST 100 index with over 95% accuracy rate. This implies that artificial neural networks provide a promising substitute for stock market predictions for economists and practitioners.



Keywords
Stock returns, Artificial Neural Networks, Istanbul Stock Exchange, Predicting.