svd, pca - linear transformation

1 回表示 (過去 30 日間)

sas0701 2014 年 3 月 13 日

編集済み: Walter Roberson 2014 年 3 月 16 日

Hi, I used to do the following..

[U,SS,V] = svd(Y*Y');
W = sqrt(SS);
w = diag(W);
L = diag(1./w)*U';
T = L*Y;

Here the first row of L gave me the linear transformation between my observations and the first PC.

FOr a variety of reasons I am now using

[coeffZ, score, latent, tsquared, explained, mu]=pca(Z);

Can someone please explain how I can obtain L from this?

Thanks,

sas0701 2014 年 3 月 16 日

anyone?

Help Center および File Exchange で Dimensionality Reduction and Feature Extraction についてさらに検索

Find the treasures in MATLAB Central and discover how the community can help you!

Translated by