How do i obtain only the first principal component?
13 visualizzazioni (ultimi 30 giorni)
Jan il 23 Ott 2013
Modificato: Jan il 23 Ott 2013
I'm not sure, if I fully understand your question. I doubt however, that there is a straightforward method for calculating the eigenvector corresponding to the largest eigenvalue of the covariance matrix without calculating all eigenvalues first (at least not for non-sparse matrices).
If you want the first principal component of the (m x n)-matrix A containing m measurements as row vectors you would in general do the following:
A = randn(100, 20); % artificial sample matrix
c_A = cov(A);
[V, ~], eigs( c_A );
p_1 = V( :, 1 );
which gives you the direction of the first principal component in the variable p_1.
Più risposte (1)
Andrew Knyazev il 12 Ago 2018
https://www.mathworks.com/matlabcentral/fileexchange/48-lobpcg-m can be used as the method for calculating the eigenvector corresponding to the largest eigenvalue of the covariance matrix without calculating all eigenvalues, or even without explicitly calculating the covariance matrix itself.