Skip to content

Commit 4162ade

Browse files
authored
DOC Update changelog before the 1.2.1 release (#25436)
1 parent c3340c0 commit 4162ade

File tree

2 files changed

+34
-26
lines changed

2 files changed

+34
-26
lines changed

doc/whats_new/v1.2.rst

Lines changed: 34 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -22,6 +22,20 @@ random sampling procedures.
2222
into account.
2323
:pr:`25354` by :user:`Jérémie du Boisberranger <jeremiedbb>`.
2424

25+
- |Fix| The `categories_` attribute of :class:`preprocessing.OneHotEncoder` now
26+
always contains an array of `object`s when using predefined categories that
27+
are strings. Predefined categories encoded as bytes will no longer work
28+
with `X` encoded as strings. :pr:`25174` by :user:`Tim Head <betatim>`.
29+
30+
Changes impacting all modules
31+
-----------------------------
32+
33+
- |Fix| Support `pandas.Int64` dtyped `y` for classifiers and regressors.
34+
:pr:`25089` by :user:`Tim Head <betatim>`.
35+
36+
- |Fix| Remove spurious warnings for estimators internally using neighbors search methods.
37+
:pr:`25129` by :user:`Julien Jerphanion <jjerphan>`.
38+
2539
Changelog
2640
---------
2741

@@ -64,6 +78,12 @@ Changelog
6478
and :class:`ensemble.ExtraTreesRegressor` now support sparse readonly datasets.
6579
:pr:`25341` by :user:`Julien Jerphanion <jjerphan>`
6680

81+
:mod:`sklearn.feature_extraction`
82+
.................................
83+
84+
- |Fix| :class:`feature_extraction.FeatureHasher` raises an informative error
85+
when the input is a list of strings. :pr:`25094` by `Thomas Fan`_.
86+
6787
:mod:`sklearn.linear_model`
6888
...........................
6989

@@ -78,6 +98,20 @@ Changelog
7898
- |Fix| :class:`manifold.TSNE` now works correctly when output type is
7999
set to pandas :pr:`25370` by :user:`Tim Head <betatim>`.
80100

101+
:mod:`sklearn.model_selection`
102+
..............................
103+
104+
- |Fix| :func:`model_selection.cross_validate` with multimetric scoring in
105+
case of some failing scorers the non-failing scorers now returns proper
106+
scores instead of `error_score` values.
107+
:pr:`23101` by :user:`András Simon <simonandras>` and `Thomas Fan`_.
108+
109+
:mod:`sklearn.neural_network`
110+
.............................
111+
112+
- |Fix| :class:`neural_network.MLPClassifier` and :class:`neural_network.MLPRegressor`
113+
no longer raise warnings when fitting data with feature names.
114+
:pr:`24873` by :user:`Tim Head <betatim>`.
81115

82116
:mod:`sklearn.preprocessing`
83117
............................

doc/whats_new/v1.3.rst

Lines changed: 0 additions & 26 deletions
Original file line numberDiff line numberDiff line change
@@ -29,11 +29,6 @@ random sampling procedures.
2929
`transform_algorithm` is not the same as `fit_algorithm` and the number of iterations
3030
is small. :pr:`24871` by :user:`Omar Salman <OmarManzoor>`.
3131

32-
- |Fix| The `categories_` attribute of :class:`preprocessing.OneHotEncoder` now
33-
always contains an array of `object`s when using predefined categories that
34-
are strings. Predefined categories encoded as bytes will no longer work
35-
with `X` encoded as strings. :pr:`25174` by :user:`Tim Head <betatim>`.
36-
3732
Changes impacting all modules
3833
-----------------------------
3934

@@ -123,34 +118,13 @@ Changelog
123118
- |Enhancement| Added the parameter `fill_value` to :class:`impute.IterativeImputer`.
124119
:pr:`25232` by :user:`Thijs van Weezel <ValueInvestorThijs>`.
125120

126-
:mod:`sklearn.feature_extraction`
127-
.................................
128-
129-
- |Fix| :class:`feature_extraction.FeatureHasher` raises an informative error
130-
when the input is a list of strings. :pr:`25094` by `Thomas Fan`_.
131-
132-
:mod:`sklearn.model_selection`
133-
..............................
134-
135-
- |Fix| :func:`model_selection.cross_validate` with multimetric scoring in
136-
case of some failing scorers the non-failing scorers now returns proper
137-
scores instead of `error_score` values.
138-
:pr:`23101` by :user:`András Simon <simonandras>` and `Thomas Fan`_.
139-
140121
:mod:`sklearn.naive_bayes`
141122
..........................
142123

143124
- |Fix| :class:`naive_bayes.GaussianNB` does not raise anymore a `ZeroDivisionError`
144125
when the provided `sample_weight` reduces the problem to a single class in `fit`.
145126
:pr:`24140` by :user:`Jonathan Ohayon <Johayon>` and :user:`Chiara Marmo <cmarmo>`.
146127

147-
:mod:`sklearn.neural_network`
148-
.............................
149-
150-
- |Fix| :class:`neural_network.MLPClassifier` and :class:`neural_network.MLPRegressor`
151-
no longer raise warnings when fitting data with feature names.
152-
:pr:`24873` by :user:`Tim Head <betatim>`.
153-
154128
:mod:`sklearn.pipeline`
155129
.......................
156130

0 commit comments

Comments
 (0)