@@ -79,13 +79,6 @@ more details.
79
79
:class: `model_selection.GridSearchCV ` object or the underlying scorer.
80
80
:pr: `27560 ` by :user: `Omar Salman <OmarManzoor> `.
81
81
82
- Metadata Routing
83
- ----------------
84
-
85
- The following models now support metadata routing in one or more or their
86
- methods. Refer to the :ref: `Metadata Routing User Guide <metadata_routing >` for
87
- more details.
88
-
89
82
- |Feature | :class: `linear_model.RANSACRegressor ` now supports metadata routing
90
83
in its ``fit ``, ``score `` and ``predict `` methods and route metadata to its
91
84
underlying estimator's' ``fit ``, ``score `` and ``predict `` methods.
@@ -101,6 +94,13 @@ more details.
101
94
transformers' ``fit `` and ``fit_transform ``. :pr: `28205 ` by :user: `Stefanie
102
95
Senger <StefanieSenger> `.
103
96
97
+ - |Fix | Fix an issue when resolving default routing requests set via class
98
+ attributes.
99
+ :pr: `28435 ` by `Adrin Jalali `_.
100
+
101
+ - |Fix | Fix an issue when `set_{method}_request ` methods are used as unbound
102
+ methods, which can happen if one tries to decorate them.
103
+ :pr: `28651 ` by `Adrin Jalali `_.
104
104
105
105
Changelog
106
106
---------
@@ -277,6 +277,14 @@ Changelog
277
277
- |Feature | :func: `sklearn.metrics.check_scoring ` now returns a multi-metric scorer
278
278
when `scoring ` as a `dict `, `set `, `tuple `, or `list `. :pr: `28360 ` by `Thomas Fan `_.
279
279
280
+ - |Fix |:class: `metrics.classification_report ` now shows only accuracy and not
281
+ micro-average when input is a subset of labels.
282
+ :pr: `28399 ` by :user: `Vineet Joshi <vjoshi253> `.
283
+
284
+ - |Fix | Fix OpenBLAS 0.3.26 dead-lock on Windows in pairwise distances
285
+ computation. This is likely to affect neighbor-based algorithms.
286
+ :pr: `28692 ` by :user: `Loïc Estève <lesteve> `.
287
+
280
288
- |API | :func: `metrics.precision_recall_curve ` deprecated the keyword argument `probas_pred `
281
289
in favor of `y_score `. `probas_pred ` will be removed in version 1.7.
282
290
:pr: `28092 ` by :user: `Adam Li <adam2392> `.
@@ -285,6 +293,14 @@ Changelog
285
293
in favor of `y_proba `. `y_prob ` will be removed in version 1.7.
286
294
:pr: `28092 ` by :user: `Adam Li <adam2392> `.
287
295
296
+ :mod: `sklearn.mixture `
297
+ ......................
298
+
299
+ - |Fix | The `converged_ ` attribute of :class: `mixture.GaussianMixture ` and
300
+ :class: `mixture.BayesianGaussianMixture ` now reflects the convergence status of
301
+ the best fit whereas it was previously `True ` if any of the fits converged.
302
+ :pr: `26837 ` by :user: `Krsto Proroković <krstopro> `.
303
+
288
304
:mod: `sklearn.model_selection `
289
305
..............................
290
306
@@ -299,13 +315,19 @@ Changelog
299
315
implementations of the Array API like CuPY.
300
316
:pr: `28407 ` by `Tim Head <betatim> `.
301
317
302
-
303
318
:mod: `sklearn.multioutput `
304
319
..........................
305
320
306
- - |Enhancement | `chain_method ` parameter added to `:class: `` multioutput.ClassifierChain`.
321
+ - |Enhancement | `chain_method ` parameter added to `:class: `multioutput.ClassifierChain `.
307
322
:pr: `27700 ` by :user: `Lucy Liu <lucyleeow> `.
308
323
324
+ :mod: `sklearn.neighbors `
325
+ ........................
326
+
327
+ - |Fix | Fixes :class: `neighbors.NeighborhoodComponentsAnalysis ` such that
328
+ `get_feature_names_out ` returns the correct number of feature names.
329
+ :pr: `28306 ` by :user: `Brendan Lu <brendanlu> `.
330
+
309
331
:mod: `sklearn.pipeline `
310
332
.......................
311
333
@@ -341,6 +363,10 @@ Changelog
341
363
- |API | :func: `utils.tosequence ` is deprecated and will be removed in version 1.7.
342
364
:pr: `28763 ` by :user: `Jérémie du Boisberranger <jeremiedbb> `.
343
365
366
+ - |Fix | :func: `~utils._safe_indexing ` now works correctly for polars DataFrame when
367
+ `axis=0 ` and supports indexing polars Series.
368
+ :pr: `28521 ` by :user: `Yao Xiao <Charlie-XIAO> `.
369
+
344
370
.. rubric :: Code and documentation contributors
345
371
346
372
Thanks to everyone who has contributed to the maintenance and improvement of
0 commit comments