@@ -279,16 +279,13 @@ Changelog
279
279
specific shape for `coef_ ` (e.g. :class: `feature_selection.RFE `).
280
280
:pr: `22016 ` by :user: `Guillaume Lemaitre <glemaitre> `.
281
281
282
- <<<<<<< HEAD
283
282
- |API | add the fitted attribute `intercept_ ` to
284
283
:class: `cross_decomposition.PLSCanonical `,
285
284
:class: `cross_decomposition.PLSRegression `, and
286
285
:class: `cross_decomposition.CCA `. The method `predict ` is indeed equivalent to
287
286
`Y = X @ coef_ + intercept_ `.
288
287
:pr: `22015 ` by :user: `Guillaume Lemaitre <glemaitre> `.
289
288
290
- =======
291
- >>>>>>> 4942ba76eb (MAINT Sort and clean-up whats new 1.1 (#23216))
292
289
:mod: `sklearn.datasets `
293
290
.......................
294
291
@@ -500,12 +497,9 @@ Changelog
500
497
`warm_start ` enabled.
501
498
:pr: `22106 ` by :user: `Pieter Gijsbers <PGijsbers> `.
502
499
503
- <<<<<<< HEAD
504
500
- |Efficiency | Improve runtime performance of :class: `ensemble.IsolationForest `
505
501
by skipping repetitive input checks. :pr: `23149 ` by :user: `Zhehao Liu <MaxwellLZH> `.
506
502
507
- =======
508
- >>>>>>> 4942ba76eb (MAINT Sort and clean-up whats new 1.1 (#23216))
509
503
- |Fix | Change the parameter `validation_fraction ` in
510
504
:class: `ensemble.GradientBoostingClassifier ` and
511
505
:class: `ensemble.GradientBoostingRegressor ` so that an error is raised if anything
@@ -774,13 +768,10 @@ Changelog
774
768
of sample weights when the input is sparse.
775
769
:pr: `22899 ` by :user: `Jérémie du Boisberranger <jeremiedbb> `.
776
770
777
- <<<<<<< HEAD
778
771
- |Fix | :class: `linear_model.SGDRegressor ` and :class: `linear_model.SGDClassifier ` now
779
772
computes the validation error correctly when early stopping is enabled.
780
773
:pr: `23256 ` by :user: `Zhehao Liu <MaxwellLZH> `.
781
774
782
- =======
783
- >>>>>>> 4942ba76eb (MAINT Sort and clean-up whats new 1.1 (#23216))
784
775
- |API | :class: `linear_model.LassoLarsIC ` now exposes `noise_variance ` as
785
776
a parameter in order to provide an estimate of the noise variance.
786
777
This is particularly relevant when `n_features > n_samples ` and the
0 commit comments