|
When working with N x N Toeplitz matrices fast algorithms (less than O(N^3)) exist for taking the inverse and back substitution (e.g. Levinson-Durbin and Trench). Is there anything faster than cubic for taking the Cholesky decomposition of a Toeplitz matrix? |
|
Supposedly there's an O(N^2) algorithm from Bareiss (1969) for the Cholesky factorization of Toeplitz matrices. |
