基于小波包去噪的股价组合预测模型

he purpose of this article is to use the hybrid model based on the wavelet packet denoising ,BP model, ARMA model and ES model model to examaine the effectiveness of forecasting stock price.The Selection data is China construction bank in 2010-2012, three years of daily closing price. First,with the original data of construction bank, three single models, BP model, ARMA model, ES model are set up respectively, then using particle swarm optimization algorithm combine the weight of models.It finds that the effectiveness in predicting the stock price of the hybrid model is better than any other single model. Then with the data after wavelet packet denoising of China construction bank ,the same hybrid model is set up.According to the results, based on the wavelet packet denoising model's prediction effectiveness is more better than the unwavelet packet denoising model.Thereby the hybrid model based on wavelet packet denoising is effective for stock price forecasting.