Presenting a Predictive Model of Stock and Over-the-Counter Market Price Indices Based on Macroeconomic Indicators Using Artificial Neural Networks

Abstract

The primary objective of this study is to present a predictive model for stock and over-the-counter (OTC) market price indices based on the examination and prediction of the effects of macroeconomic variables on Iran’s capital market. To improve prediction accuracy, the study employed deep neural network algorithms to forecast future trends in stock price indices. In this research, the required data were collected from the Central Bank of Iran, the Statistical Center of Iran, the Gold and Currency Information Website, and the Securities and Exchange Organization (SEO) database for stock price index data. The analysis was conducted using two artificial neural network approaches: the Long Short-Term Memory (LSTM) network and the Multilayer Perceptron (MLP) trained through the backpropagation algorithm. The models were evaluated using Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and prediction accuracy percentage. The implementation was carried out in the Python programming environment over the period from 2014 to 2023. The findings of the study indicated that the LSTM neural network model did not yield satisfactory prediction results. However, the MLP model successfully demonstrated the predictive impact of macroeconomic variables on the stock and OTC market price indices. Overall, the results suggest that the application of artificial neural networks, as an advanced analytical tool, provides an effective and efficient approach for analyzing and predicting stock market fluctuations. This study emphasizes the importance of considering macroeconomic variables in the analysis and forecasting of stock market volatility. Furthermore, the findings can serve as a valuable guide for economic policymakers and investors in making optimal decisions within financial markets.