Full Length Research Paper
Abstract
Stocks return prediction is one of the most important discussions of financial analysts in recent decades. The majority of researchers including Balvers et al. (1990), Breen et al. (1990), Campbell (1987), Fama (1977), Ferson (1989), Keim and Stambaugh (1986) and Schwert (1990) showed evidences that stocks return can be predicted by means of historical information such as periodic data of financial and economic variables. Although most studies indicated that the relationship between available information and stocks return is based on linear assumptions, there are no evidences that this relationship is completely linear. Therefore, it is probable that non-linear models have more ability to predict stocks price fluctuations and their returns. The present paper predicted the stocks return with artificial neural network (ANN) and linear regression in Tehran Stocks Exchange and New York Stocks Exchange (NYSE) using financial and macroeconomic variables. Furthermore, the results of two markets were compared. As it is known, stocks return depends on many variables. Thus, in this research, variables were screened and irrelevant or redundant variables were removed using principal component analysis (PCA). Then, in order to predict the stocks return, feed-forward and linear regression model were applied. Based on the mean square error (MSE), the results showed that the estimated error of ANN in two markets is less than the estimated error of linear regression. In addition, ANN and linear regression model will make a better prediction in NYSE.
Key words: Artificial neural network, feed-forward, financial market, principal component analysis, regression analysis.
Copyright © 2024 Author(s) retain the copyright of this article.
This article is published under the terms of the Creative Commons Attribution License 4.0