num_iden_gross0.m

基于GM算法和QR分解实现的稳健奇异值分解算法

function [t,scale]=num_iden_gross0(y)
%Identify the gross errors'numbers of an observational vector y
% 
t=zeros(1);
n=length(y);% Compute the size of the vector y

mean_y=mean(y);%Compute the mean of the vector y
v=y-mean_y;% Compute the residuas of the vector y
scale=median(v.^2)^0.5*1.4826*(1+5/(n-1));% Compute the robust scale parameter, i.e. the standard error, of the vector y
m=0;
for i=1:n 
    m=m+1;
    if abs(v(i))>scale*2.5
        t(m)=i;% Identify the location number of the gross error; 
    else
        t(m)=0;
    end    
end
end