ssa.asv

不同算法对时间序列进行奇异谱分解的源程序代码

  function [E,V,A,R,C]=ssa(x,M,method) 
%  SSA - driver routine to perform Singular Spectrum Analysis
%  Syntax: [E,V,A,R,C]=ssa(x,M); [E,V,A,R,C]=ssa(x,M,'BK');
%
% Input:    x - time series
%           M - embedding dimension.   嵌入维数
%      method - (optional) method of calculating the covariance matrix:
%                 'unbiased' (N-k weighted) (default)     无偏的
%                 'biased'   (N-weighted or Yule-Walker)  有偏的
%                 'BK'       (Broomhead/King type estimate) ？？ 
%
% Output: E - eigenfunction (T-EOF) matrix in standard form   标准形式的时间的EOF
%         V - vector containing variances (unnormalized eigenvalues)  向量包含变量（非标准形式的特征向量）   
%         A - Matrix of principal components 主成分矩阵
%         R - Matrix of reconstructed components   重建成分矩阵
%         C - Covariance matrix 协方差矩阵
%
%  See Vautard, Yiou, and Ghil, Physica D 58, 95-126, 1992.
%
%  Written by Eric Breitenberger.     Version date 1/22/96
%  Please send comments and suggestions to [e-mail]   

if nargin==2, method=''; end

[E,V,C]=ssaeig(x,M,method);
[A]=pc(x,E);
[R]=rc(A,E);

unbiased