在线阅读 --自然科学版 2014年1期《基于EEMD和改进Elman神经网络的地球变化磁场短时预测》
基于EEMD和改进Elman神经网络的地球变化磁场短时预测--[在线阅读]
牛超, 卢世坤, 祁树锋
第二炮兵工程大学 907教研室, 陕西 西安 710025
起止页码: 50--54页
DOI: 10.11826/j.issn.1000-5854.2014.01.011
摘要
针对地球变化磁场时间序列的混沌特性,提出了一种集成经验模态分解(ensemble empirical mode decomposition,EEMD)和改进Elman神经网络的地球变化磁场预测模型.首先,利用EEMD将非平稳的地球变化磁场时间序列分解为一系列具有不同特征尺度的本征模态函数(intrinsic mode function,IMF);然后,针对每一个IMF分别建立改进Elman神经网络模型,选择各自适合的最优模型参数;最后,以地磁台站实测的地球变化磁场数据为研究对象,并与基于单一Elman神经网络预测模型相比较,结果表明,EEMD 改进Elman神经网络模型的预测值能紧跟地球变化磁场的变化趋势,且明显优于基于单一Elman神经网络的模型,体现出更好的预测效果.在地磁Kp<3时,预测3h平均绝对误差为1.74nT.

Prediction of the Geomagnetic Variation Field Based on Modified Ensemble Empirical Mode Decomposition and Modified Elman Neural Network
NIU Chao, LU Shikun, QI Shufeng
907 Department, the Second Artillery Engineering University, Shaanxi Xi'an 710025, China
Abstract:
According to the chaotic feature of geomagnetic variation time series, a combined forecasting model based on ensemble empirical mode decomposition (EEMD)and modified Elman neural network is proposed.Firstly, the geomagnetic variation time series is decomposed into a series of intrinsic mode function (IMF)with different characteristic scales by using EEMD.Then, the forecasting model of each IMF is created with modified Elman neural network, through using the optimal model parameters.Finally, the simulation is performed by using the real data collected from the geomagnetic observatory, and its results were compared with the model based on single Elman neural network.The results show that the forecasting value of the EEMD-modified Elman neural network model can closely keep up with the trend of geomagnetic variation field obviously better than the model based on single Elman neural network.The mean absolute error of the model forecasting three hours is 1.74 nT when Kp less than 3.

收稿日期: 2013-6-20
基金项目: 国家自然科学基金(40974037)

参考文献:
[1]徐文耀.地球电磁现象物理学 [M].合肥:中国科学技术大学出版社,2009:473-474.
[2]徐文耀.地磁活动K指数值量算和确定方法的改进 [J].西北地震学报,2005,27(增):36-41.
[3]彭富清.地磁模型与地磁导航 [J].海洋测绘,2006,26(2):73-75.
[4]ELMAN J L.Finding Structure in Time [J].Cognitive Sci,1990,14:179-211.
[5]HUANG N E,SHEN Z,LONG S R,et al.The Empirical Mode Decomposition and the Hilbert Spectrum for Nonlinear and Non-startionary Time Series Analysis [J].Proc Roy Soc,1998,454A:903-995.
[6]WU Z H,HUANG N E.Ensemble Empirical Mode Decomposition:A Noise Assisted Data Analysis Method [J].Advances in Adaptive Data Analysis,2009,1(1):1-41.
[7]余健,郭平.基于改进的Elman神经网络的股价预测模型 [J].计算机技术与发展,2008,18(3):43-45.
[8]齐玮,王秀芳,李夕海,等.基于统计建模的地磁匹配特征量选择 [J].地球物理学进展,2010,25(1):324-330.
[9]CHEN G X,XU W Y,DU A M.Statistical Characteristics of the Day-to-day Variability in the Geomagnetic Sq Field [J].J Geophys Res,2007,112:A06320.