ENH: pseudo-determinant, matrix determinant for singular cov, mahalanobis

[josef-pkt]

https://documentation.sas.com/doc/en/statcdc/14.2/statug/statug_rreg_details26.htm#statug.rreg.robustreggrd

SAS MCD takes into account that cov in subset could be singular.
Standard MCD algorithm assumes full rank.

pseudo determinant is product of non-zero, strictly positive eigenvalues (where numeric cutoff is defined by a condition number threshold, but I don't find want cutoff value they use)

related: singular cov in mahalanobis distance
We can use pinv for in-sample distances.
But when evaluating out-of-sample observations, then points might not be in the same subspace.