摘要
金融时间序列数据的预测是预测领域的热点问题。本文结合小波变换与神经网络的有关理论,给出了基于小波神经网络的石油期货价格预测具体学习算法并进行了拟合及检测,结果表明该方法具有比常用的BP算法及径向基函数网络算法(HCM算法)更好的拟合能力、推广能力,可为石油期货买卖决策提供一定的依据,并可推广于其它金融时间序列的预测。
The forecasting of finance time series data is hot spot questions in forecasting fields. This paper gives out the learning algorithm of forecasting about petroleum futures price by combining the wavelet transform and neural network theory. The learning and testing results prove wavelet neural network forecasting algorithm superior to BP algorithm and radial basis function network algorithm (HCM algorithm) in common use. So, wavelet neural network forecasting method can provide some basis for the petroleum futures buying and sellinges, and can popularize in other finance time series forecasting.
出处
《技术经济》
2006年第6期121-124,共4页
Journal of Technology Economics
基金
福州大学科技发展基金资助(2004-XQ(S)-05)
关键词
小波神经网络
石油期货价格
预测
wavelet neural network
petroleum futures price
forecasting