Open Shuyib opened 6 years ago
I took a stab at the Lasso and ElasticNet classifiers. LogisticRegression looks like a pain, because each solver has different restrictions on the input and other parameters. It's doable, but I wonder if it might make more sense to have a component for each solver, and then provide a function (like any_classifier
) that chooses between them.
Sounds like a fantastic plan.
Hi,
I've run hyperopt-sklearn successfully and i really like it. Is it possible to add Logistic Regression, ElasticNet and Lasso regression? I find them a bit difficult to tune sometimes. Thanks!