Applying matrix operations to `Fit`

[odunbar]

LinearAlgebra defines operations such as det() on Factorization types such as cholesky etc.
currently to calculate this, we need a tree such as

det(get_decomposition(get_feature_factors(fit::Fit)))

It would be nice to reduce this to det(fit::Fit) and returning sensible options such as

• In the applicable factorizations, the usual determinant
• In the pseudo inverse case, the det of the original matrix,
• In the svd factorization perhaps the product of singular values (note, this gives only the abs(det(fit)) - presumable why it is not in the list of applicable factorizations by LinearAlgebra)
  etc.

PS: The determinant is so far the only operation I've needed but perhaps there might be other useful ones.