Artificial neural networks : A tool for forecasting stock returns

Abstract

This article examines and analyzes the use of neural networks as a forecasting tool. Specifically an artificial neural network's ability, to predict future trend of prices of stocks against actual price, included in the Sensitive Index (Sensex) of Bombay Stock Exchange (BSE), is tested. Scope of this study is restricted to individual investor. In this study, authors have used 11 proves to forecast the stock returns. Results of the study are encouraging and the deviations are less than 5 percent. During the training session, it is observed that not all the probes used are relevant for all the stocks and therefore, it is necessary to identify the probes, which are influencing the specific stock. Artificial neural network is not a black box, if it is used judiciously, the results could be amazing. The findings suggest that stock markets do not follow a random walk and there exists a possibility of predicting stock returns. Authors opine that it is possible to capture non-linearities contained in the stock returns by using artificial neural network. If neural network is used astutely, it could benefit the individual investors.