摘要
研究股价预测问题,针对股价指数具有不稳定和时变性,单一预测方法预测准确度低、误差过大,为提高预测精度,消除噪声,提出一种小波分析的自回归滑动平均(ARIMA)与BP神经网络(BPNN)相结合的股价指数组合预测方法。组合预测方法首先采用小波分析对股价原始数据进行分解和重构,股价数据经过小波处理后,变成线性和非线性两部分,采用ARIMA和BPNN分别对线性部分和非线部分进行预测,最后组合两者预测结果得到股价指数最终预测结果,用上证A股的收盘指数数据对组合预测方法进行了验证测试,实验结果表明组合预测方法比单一预测方法预测准确度高,误差小,在股价指数预测中具有广泛的应用前景,可为股市提供参考。
Based on the analysis of the characteristics of nonlinearity and strong interference of stock price due to the complex and uncertainty of time variance,a new approach was proposed for stock price prediction.Firstly,wavelet transform is employed to decomposition the original stock price data to reflect the essence and variation of t stock price Then a hybrid methodology that exploits the unique strength of the ARIMA model and BPNN model to forecast t stock price Finally,numerical field examples were given to testify the precision of the model.The result shows that（1） the hybrid model can produce more accurate predictions than that of single model;（2） the hybrid model that uses the method of wavelet decomposition is more efficient and reliable.The hybrid model based on wavelet decomposition can be an efficient method to the dynamic stock price prediction.
出处
《计算机仿真》
CSCD
北大核心
2010年第10期297-300,共4页
Computer Simulation
关键词
组合预测
小波分析
神经网络
股票价格
Combination forecast
Wavelet analysis
Neural network
Stock price
