Forecasting opening stock prices by integrating signal processing and the artificial neural network: an exploratory study

Abstract

This study applies the use of an integrated signal processing technique with artificial neural network to evaluate its effectiveness in predicting individual stock pricing. The technique was previously used to predict stock market behaviour with an effective average prediction accuracy of 96.7%. We apply it to individual stocks to determine whether the model has similar predictive accuracy for individual stocks and to determine whether the accuracy matters based on the longevity of the stock availability. Archived data of Kohls, JCPenneys, Apple and Blackberry were used for training and testing the proposed model. The results strongly support the effectiveness of the proposed model with an overall average prediction accuracy of 79%. This accuracy suggests that with additional study, a better model may be possible. This exploratory study expands current forecasting research by applying a recent stock market forecasting method to individual stock price forecasting.