1- Department of Economics and Finance, Abhar Branch, Islamic Azad University, Abhar, Iran 2- Department of Economics and Finance, Abhar Branch, Islamic Azad University, Abhar, Iran , 4410247239@iau.ac.ir 3- Department of Economics and Finance, Zanjan Branch, Islamic Azad University, Zanjan, Iran 4- Department of Economic Sciences, University of Mohaghegh Ardabili, Ardabil, Iran
Abstract: (13 Views)
Financial markets, due to their inherent characteristics such as complexity, dynamics, and dependence on various economic and social factors, require advanced analytical methods to understand trends and predict future behaviors. Therefore, this study was designed with the aim of providing a model for predicting stock exchanges using macroeconomic variables and social network data using linear lasso methods and nonlinear Gaussian process. In this study, information on macroeconomic variables and social networks was examined for 15 stocks listed on the Tehran Stock Exchange between 1396 and 1402. The results of this study showed that the combination of macroeconomic variables including interest rates, inflation rates, exchange rates, gross domestic product, and social network data significantly improves the accuracy of predicting stock exchanges. It was also found that the power of "machine learning" methods for predicting stock exchanges is greater than the linear lasso method. The knowledge contribution of this research, in addition to designing a stock exchange prediction model and evaluating the power of machine learning methods, is within the scope of the richness of the literature in this field.
aghapourAlishahi M, Askari F, Rajaee Y, Hashemi Dizaj A. Designing a Stock Exchange Forecasting Model Using Macroeconomic Variables and Social Networks. mieaoi 2026; 15 (56) : 22 URL: http://mieaoi.ir/article-1-2000-en.html