This PR uses Solve to replace matrix inversion when calculating \hat{\beta}.
Note: Cov(\hat{\beta}) is still calculated using blm_inverse as this was just a simple case of an inverse matrix multiplied by a constant and couldn't be improved with solve. Similarly c'(X'X)^(-1)c is also computed using blm_inverse as there is no restriction on c to be full rank.
This PR uses Solve to replace matrix inversion when calculating \hat{\beta}.
Note: Cov(\hat{\beta}) is still calculated using
blm_inverseas this was just a simple case of an inverse matrix multiplied by a constant and couldn't be improved with solve. Similarly c'(X'X)^(-1)c is also computed usingblm_inverseas there is no restriction on c to be full rank.