@@ -71,7 +71,7 @@ more details.
71
71
:class: `ensemble.StackingRegressor ` now support metadata routing and pass
72
72
``**fit_params `` to the underlying estimators via their `fit ` methods.
73
73
:pr: `28701 ` by :user: `Stefanie Senger <StefanieSenger> `.
74
-
74
+
75
75
- |Feature | :class: `compose.TransformedTargetRegressor ` now supports metadata
76
76
routing in its `fit ` and `predict ` methods and routes the corresponding
77
77
params to the underlying regressor.
@@ -81,6 +81,14 @@ more details.
81
81
the `fit ` method of its estimator and for its underlying CV splitter and scorer.
82
82
:pr: `29329 ` by :user: `Stefanie Senger <StefanieSenger> `.
83
83
84
+ - |Feature | :class: `semi_supervised.SelfTrainingClassifier `
85
+ now supports metadata routing. The fit method now accepts ``**fit_params ``
86
+ which are passed to the underlying estimators via their `fit ` methods.
87
+ In addition, the `predict `, `predict_proba `, `predict_log_proba `, `score `
88
+ and `decision_function ` methods also accept ``**params `` which are
89
+ passed to the underlying estimators via their respective methods.
90
+ :pr: `28494 ` by :user: `Adam Li <adam2392> `.
91
+
84
92
Dropping official support for PyPy
85
93
----------------------------------
86
94
@@ -189,6 +197,13 @@ Changelog
189
197
when duplicate values in the training data lead to inaccurate outlier detection.
190
198
:pr: `28773 ` by :user: `Henrique Caroço <HenriqueProj> `.
191
199
200
+ :mod: `sklearn.semi_supervised `
201
+ ..............................
202
+
203
+ - |API | :class: `semi_supervised.SelfTrainingClassifier `
204
+ deprecated the `base_estimator ` parameter in favor of `estimator `.
205
+ :pr: `28494 ` by :user: `Adam Li <adam2392> `.
206
+
192
207
Thanks to everyone who has contributed to the maintenance and improvement of
193
208
the project since version 1.5, including:
194
209
0 commit comments