博文
Theil Sen Trend Estimator
|
function b = get_trend(y)
% b is the slope, i.e., changes per step.
N=size(y,1); % [col, 1]
x=(1:N)'; % [col, 1]
bls = regress(y,[ones(N,1) x]);
Comb = combnk(1:N,2);
theil=diff(y(Comb),1,2)./diff(x(Comb),1,2);
b=median(theil);
https://blog.sciencenet.cn/blog-569118-724636.html
上一篇：inpaint data gaps of N dimensions
下一篇：Mann-Kendall Tau (aka Tau-b) with Sen's Method
