Computes \(\sum_i \log(\lambda_i)\) over eigenvalues \(\lambda_i\)
of a symmetric matrix Q that exceed eigen_tol.
generalized_logdet(Q, eigen_tol = NULL, rank_deficiency = 0)
Arguments
- Q
Symmetric matrix.
- eigen_tol
Nonnegative threshold; eigenvalues <= eigen_tol are ignored.
If NULL, uses a heuristic based on machine precision.
- rank_deficiency
Optional integer. If > 0, drops the smallest
rank_deficiency eigenvalues (useful when you know the null-space
dimension a priori).
Value
Scalar generalized log-determinant.