摘要
针对总电子含量TEC非线性、非平稳的特性,将SSA方法引入到TEC预报中,对SSA分解并筛选合并后的分量进行ARMA模型预测,然后叠加各个分量的预报值。采用IGS提供的2010年电离层TEC数据进行实验,结果表明,利用该组合方法预报5 d的TEC的平均相对精度为92%,比仅用ARMA模型提高4%。
Based on the nonlinear and non-stationary characteristics of total electric contents time series, the combination of SSA method and ARMA model was proposed to predict TEC. In the combination, ARMA model is used to predict the data which had been decomposed and filtered by SSA method,and the predicted value of each component is overlaid. The results calculated by the combination for 2010TEC data show that the mean relative pre- cision of TEC prediction for 5 days is 92% by the combination, and the accuracy is improved of 4% , comparing with individual ARMA method.
出处
《大地测量与地球动力学》
CSCD
北大核心
2014年第6期44-49,共6页
Journal of Geodesy and Geodynamics
基金
国土环境与灾害监测国家测绘地理信息局重点实验室开放基金项目(LEDM2013B02)
湖南省国土资源厅科技项目(2012-41)
关键词
总电子含量
奇异谱分析
ARMA模型
total electric contents
singular spectrum analysis
ARMA model
