AdaBoost Artificial Neural Network for Stock Market Predicting

Abstract

In this work, we propose a new direction of stock index movement prediction algorithm, coined the Ada-ANN forecasting model, which exploits AdaBoost theory and ANN to fulfill the predicting task. ANNs are employed as the weak forecasting machines to construct one strong forecaster. Technical indicators from Chinese stock market and international stock markets such as S&P 500, NSADAQ, and DJIA are selected as the predicting independent variables for the period under investigation. Numerical results are compared and analyzed between strong forecasting machine and the weak one. Experimental results show that the Ada-ANN model works better than its rival for predicting direction of stock index movement.