"sparse inverse" from sparse Cholesky

I am looking for a manner to obtain, from the sparse Cholesky decomposition of sparse A, the “sparse inverse”, i.e. obtaining only the elements in A^{-1} that correspond to the non-null elements in A. This is sometimes known as the Takahashi inverse, and there is an efficient algorithm, see Functions — cholmod-extra 0.1 documentation . Any pointer? Thanks.

1 Like