Closed ValdarT closed 5 years ago
Note: I quickly checked against GLM.jl, results looked good but performance was worse.
Thanks, LGTM. @PastelBelem8 this was your PR, would you like to review this one?
Note: I quickly checked against GLM.jl, results looked good but performance was worse.
Do you know what algorithm they're using? Conjugate gradient? It'd be interesting to compare with scikit-learn, too.
Sure @cstjean I will take a look
It might also be the part where we're creating a new matrix (adding the intercepts) and assigning slices of it.
Fixed the LinearRegression API to input format n_samples x n_features.