Improved interpolation method based on singular spectrum analysis iteration and its application to missing data recovery  

作　　者：王辉赞 张韧 刘巍 王桂华 金宝刚 

机构地区：[1]Institute of Meteorology,PLA University of Science and Technology [2]School of Information Science and Technology,Southwest Jiaotong University [3]Second Institute of Oceanography,State Oceanic Administration

出　　处：《Applied Mathematics and Mechanics(English Edition)》2008年第10期1351-1361,共11页应用数学和力学（英文版）

基　　金：the State Key Program for Basic Research of China(No.2007CB816003);the Open Item of the State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics of China

摘　　要：A novel interval quartering algorithm （IQA） is proposed to overcome insufficiency of the conventional singular spectrum analysis （SSA） iterative interpolation for selecting parameters including the number of the principal components and the embedding dimension. Based on the improved SSA iterative interpolation, interpolated test and comparative analysis are carried out to the outgoing longwave radiation daily data. The results show that IQA can find globally optimal parameters to the error curve with local oscillation, and has advantage of fast computing speed. The improved interpolation method is effective in the interpolation of missing data.A novel interval quartering algorithm （IQA） is proposed to overcome insufficiency of the conventional singular spectrum analysis （SSA） iterative interpolation for selecting parameters including the number of the principal components and the embedding dimension. Based on the improved SSA iterative interpolation, interpolated test and comparative analysis are carried out to the outgoing longwave radiation daily data. The results show that IQA can find globally optimal parameters to the error curve with local oscillation, and has advantage of fast computing speed. The improved interpolation method is effective in the interpolation of missing data.

关 键 词：singular spectrum analysis outgoing longwave radiation interpolation of missing data interval quartering algorithm 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]