`leverage` with sparse matrix

[kylebutts]

Hi there!

I'm wondering if it's possible to have a version of the leverage function that operates on sparse matrices?

I'm trying to compute leave-out standard errors (https://github.com/rsaggio87/LeaveOutTwoWay) for a high dimensional fixed-effect model, so loading the matrix into memory isn't feasible

[ManosPapadakis95]

We would love to deal with sparse matrices. Many have asked it before. Unfortunately the sparse matrices hide many problems (in the area of mathematics) and we don't have the time to learn and deal.
After some research, I found what you want but you have to use other packages.

Crossprod for sparse: https://stackoverflow.com/questions/71972327/sparse-matrix-as-a-result-of-crossprod-of-sparse-matrices

Mahalanobis distance for sparse: https://rdrr.io/cran/DiPs/man/maha_sparse.html