Back propagation mathematical model for stock price prediction

Due to the extremely volatile nature of financial markets, it is commonly accepted that stock price prediction is a task filled with challenges. However, in order to make profits or understand the essence of equity market, numerous market participants or researchers try to forecast stock prices using various statistical, econometric or even neural network models. In this work, we survey and compare the predictive power of five neural network models, namely, back propagation (BP) neural network, radial basis function neural network, general regression neural network, support vector machine regression (SVMR) and least squares support vector machine regression. We apply the five models to make price predictions for three individual stocks, namely, Bank of China, Vanke A and Guizhou Maotai. Adopting mean square error and average absolute percentage error as criteria, we find that BP neural network consistently and robustly outperforms the other four models. Then some theoretical and practical implications have been discussed.