Unnecessary computation of inverse

[niklasfries]

src/weighted_matrix.cpp:72:

R = inv(X.t() * Z) * X.t(); should be R = solve(X.t() * Z, X.t());

This probably doesn't matter for the amounts of data the rest of the implementation is scaled for, but it's good practice and fixed a problem I had with an ill-conditioned matrix.

[martinju]

Thanks! We are working with the scalability of the whole package, so all improvements (even minor ones like this) is very welcome :-)
I'll update this in a PR next week or so.

[AtroXWorf]

@martinju Have you had time to change the above mentioned line?

[martinju]

Nope, but fixing it with a PR now :-)