How to find more than six eigenvectors of a large matrix in matlab?

Question

I have a big matrix with size 12000x12000 and I need to find 100 eigenvectors with the highest 100 eigenvalues of that matrix (in order to perform a PCA dimension reduction).

I tried using the matlab's eig function, which returns more eigenvectors than I need takes forever since the matrix size is too big. I also noticed that there the 'eigs' function is much faster but it doesn't returns only 6 eigenvectors, which is less than what I need.

any ideas what can I do?

Thanks!

Answer 1

Pass the second, optional parameter to eigs, which controls how many eigenvectors are returned.