Include ElasticNet like regularization for SparseLogisticRegression #244
Conversation
|
Sorry @mathurinm for the misunderstanding, I tried doing it, is it alright now? |
This reverts commit 4563f22.
hoodaty
changed the title
| @@ -1035,7 +1036,7 @@ def fit(self, X, y): | |||
| max_iter=self.max_iter, max_pn_iter=self.max_epochs, tol=self.tol, | |||
| fit_intercept=self.fit_intercept, warm_start=self.warm_start, | |||
| verbose=self.verbose) | |||
| return _glm_fit(X, y, self, Logistic(), L1(self.alpha), solver) | |||
| return _glm_fit(X, y, self, Logistic(), L1_plus_L2(self.alpha,self.l1ratio), solver) | |||
There was a problem hiding this comment.
this is not PEP8 compliant, by convention you need to add a space after the comma. You can configure your editor to do it automatically : https://github.com/mathurinm/github-assignment/?tab=readme-ov-file#vscode-configuration
| @@ -1003,10 +1003,11 @@ class SparseLogisticRegression(LinearClassifierMixin, SparseCoefMixin, BaseEstim | |||
| Number of subproblems solved to reach the specified tolerance. | |||
| """ | |||
|
|
|||
| def __init__(self, alpha=1.0, tol=1e-4, max_iter=20, max_epochs=1_000, verbose=0, | |||
| def __init__(self, alpha=1.0, l1ratio=0.5, tol=1e-4, max_iter=20, max_epochs=1_000, verbose=0, | |||
There was a problem hiding this comment.
You need to add a description of this parameter in the docstring above; also, call it l1_ratio for compatibility with sklearn.
|
Yes much better like this, thank you! Can you add "Closes #231" in the first message of this PR, so that the corresponding is matched and closed automatically when this is merged? As for any new feature, you need to add a new unit test to check that the implementation is correct. |
Context of the PR
Tries to add an ElasticNet like regularization for SparseLogisticRegression
Contributions of the PR
Added an L1_plus_L2 penalty
Checks before merging PR