sirt_eigenvalues.Rd
This function computes the first \(D\) eigenvalues and eigenvectors of a symmetric positive definite matrices. The eigenvalues are computed by the Rayleigh quotient method (Lange, 2010, p. 120).
sirt_eigenvalues( X, D, maxit=200, conv=10^(-6) )
X | Symmetric matrix |
---|---|
D | Number of eigenvalues to be estimated |
maxit | Maximum number of iterations |
conv | Convergence criterion |
A list with following entries:
Vector of eigenvalues
Matrix with eigenvectors in columns
Lange, K. (2010). Numerical Analysis for Statisticians. New York: Springer.
Sigma <- diag(1,3) Sigma[ lower.tri(Sigma) ] <- Sigma[ upper.tri(Sigma) ] <- c(.4,.6,.8 ) sirt::sirt_eigenvalues(X=Sigma, D=2 ) # compare with svd function svd(Sigma)