I have a 800x800 singular (covariance) matrix and I want to find it's largest eigenvalue and eigenvector corresponding to this eigenvalue. Does anybody know wheter it is possible to do it with R?
Asked
Active
Viewed 1,827 times
0
-
see `eigen` function (e.g. `result$values[1]` and `result$vectors[,1]`) – Marc in the box May 21 '13 at 14:23
-
@Marcinthebox, it's not working for singular matrix – user2080209 May 21 '13 at 14:25
-
Ok, jumped the gun I guess. You may want to provide a small example. perhaps `svd` would find a solution where `eigen` does not. – Marc in the box May 21 '13 at 14:29
-
@Marcinthebox, can't find how to use swd for my problem – user2080209 May 21 '13 at 14:33
1 Answers
1
Here is an example of using svd for the decomposition of a covariance matrix:
a <- matrix(runif(16),4)
C <- cov(a)
res <- svd(C)
res
res$d[1] # largest singular value
res$u[,1] # largest vector ; u and v are the same
Hope that helps.
Marc in the box
- 11,769
- 4
- 47
- 97