查询结果:   张彦霞,肖清泰,徐建新,桑秀丽.基于经验模态分解的小波神经网络预测模型[J].计算机应用与软件,2016,33(10):284 - 287.
中文标题
基于经验模态分解的小波神经网络预测模型
发表栏目
算法
摘要点击数
642
英文标题
WAVELET NEURAL NETWORK PREDICTION MODEL BASED ON EMPIRICAL MODEL DECOMPOSITION
作 者
张彦霞 肖清泰 徐建新 桑秀丽 Zhang Yanxia Xiao Qingtai Xu Jianxin Sang Xiuli
作者单位
昆明理工大学质量发展研究院 云南 昆明 650093 昆明理工大学省部共建复杂有色金属资源清洁利用国家重点实验室 云南 昆明 650093    
英文单位
Quality Development Institute,Kunming University of Science and Technology,Kunming 650093,Yunnan,China State Key Laboratory of Complex Nonferrous Metal Resources Clear Utilization,Kunming University of Science and Technology,Kunming 650093,Yunnan,China    
关键词
经验模态分解 小波神经网络 BP神经网络 预测
Keywords
Empirical mode decomposition Wavelet neural networks Back propagation neural network Forecasting
基金项目
作者资料
张彦霞,硕士生,主研领域:预测模型。肖清泰,硕士生。徐建新,博士。桑秀丽,教授。 。
文章摘要
针对小波神经网络(WNN)在非平稳、非线性时间序列预测上无法实现自适应多分辨率分析,且其预测精度有待提高的问题,提出基于经验模态分解的小波神经网络预测模型。首先,对非线性、非平稳时间序列进行经验模态分解(EMD),以降低时间序列的非平稳性;然后对EMD分析得到的固有模态分量(IMF)和余项分别构建WNN模型;最后,汇总预测结果,得到预测值。通过数据验证,新模型的预测精度高于BP神经网络和WNN。
Abstract
Wavelet neural network (WNN) can’t achieve adaptive multi-resolution analysis on non-stationary and nonlinear time series prediction, and its prediction accuracy needs to be improved. In order to solve the problems above, this paper proposes an EMD-based prediction model of wavelet neural network. First, it applies empirical mode decomposition (EMD) on non-linear and non-stationary time series so as to reduce the non-stationarity of time series. Then, it builds respectively the WNN models of intrinsic mode functions (IMF) and remainders derived from EMD analysis. Finally, it summarises the results of each prediction to obtain the final forecasting value. Through data verification, it is proved that the prediction accuracy of new model is higher than that of BP neural network and WNN.
下载PDF全文