8000 MNT what's new corrections · lithuak/scikit-learn@6d9acd7 · GitHub
[go: up one dir, main page]
Skip to content

Commit 6d9acd7

Browse files
TomDLTTomDLT
committed
MNT what's new corrections
1 parent f24e300 commit 6d9acd7

File tree

2 files changed

+50
-55
lines changed

2 files changed

+50
-55
lines changed

‎doc/whats_new/v0.20.rst

Lines changed: 7 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -70,7 +70,7 @@ Changelog
7070

7171
- |Fix| Fixed a regression in :class:`decomposition.IncrementalPCA` where
7272
0.20.0 raised an error if the number of samples in the final batch for
73-
fitting IncrementalPCA was smaller than n_components.
73+
fitting IncrementalPCA was smaller than n_components.
7474
:issue:`12234` by :user:`Ming Li <minggli>`.
7575

7676
:mod:`sklearn.ensemble`
@@ -88,19 +88,18 @@ Changelog
8888
:mod:`sklearn.linear_model`
8989
...........................
9090

91-
- |Fix| :func:`linear_model.SGDClassifier` and variants
91+
- |Fix| :class:`linear_model.SGDClassifier` and variants
9292
with ``early_stopping=True`` would not use a consistent validation
9393
split in the multiclass case and this would cause a crash when using
9494
those estimators as part of parallel parameter search or cross-validation.
9595
:issue:`12122` by :user:`Olivier Grisel <ogrisel>`.
9696

97-
- |Fix| Fixed a bug affecting :func:`SGDClassifier.fit` in the multiclass
97+
- |Fix| Fixed a bug affecting :class:`SGDClassifier` in the multiclass
9898
case. Each one-versus-all step is run in a :class:`joblib.Parallel` call and
9999
mutating a common parameter, causing a segmentation fault if called within a
100100
backend using processes and not threads. We now use ``require=sharedmem``
101-
at the :class:`Parallel` instance creation.
102-
:issue:`12518` by :user:`Pierre Glaser <pierreglaser>` and
103-
:user:`Olivier Grisel <ogrisel>`.
101+
at the :class:`joblib.Parallel` instance creation. :issue:`12518` by
102+
:user:`Pierre Glaser <pierreglaser>` and :user:`Olivier Grisel <ogrisel>`.
104103

105104
:mod:`sklearn.metrics`
106105
......................
@@ -158,7 +157,7 @@ Miscellaneous
158157
- |Fix| Make sure to avoid raising ``FutureWarning`` when calling
159158
``np.vstack`` with numpy 1.16 and later (use list comprehensions
160159
instead of generator expressions in many locations of the scikit-learn
161-
code base). :issue:`12467` by :user:`Olivier Grisel`.
160+
code base). :issue:`12467` by :user:`Olivier Grisel <ogrisel>`.
162161

163162
.. _changes_0_20:
164163

@@ -1032,7 +1031,7 @@ Support for Python 3.3 has been officially dropped.
10321031
:issue:`9304` by :user:`Breno Freitas <brenolf>`.
10331032

10341033
- |API| :class:`pipeline.FeatureUnion` now supports ``'drop'`` as a transformer
1035-
to drop features. :issue:`11144` by :user:`thomasjpfan`.
1034+
to drop features. :issue:`11144` by :user:`Thomas Fan <thomasjpfan>`.
10361035

10371036

10381037
:mod:`sklearn.preprocessing`

‎doc/whats_new/v0.21.rst

Lines changed: 43 additions & 47 deletions
Original file line numberDiff line numberDiff line change
@@ -40,57 +40,20 @@ Support for Python 3.4 and below has been officially dropped.
4040
- An entry goes here
4141
- An entry goes here
4242

43-
:mod:`sklearn.linear_model`
44-
...........................
45-
- |Fix| Fixed a bug in :class:`linear_model.MultiTaskElasticNet` which was breaking
46-
``MultiTaskElasticNet`` and ``MultiTaskLasso`` when ``warm_start = True``. :issue:`12360`
47-
by :user:`Aakanksha Joshi <joaak>`.
48-
4943
:mod:`sklearn.cluster`
5044
......................
5145

5246
- |MajorFeature| A new clustering algorithm: :class:`cluster.OPTICS`: an
5347
algoritm related to :class:`cluster.DBSCAN`, that has hyperparameters easier
5448
to set and that scales better, by :user:`Shane <espg>`,
5549
:user:`Adrin Jalali <adrinjalali>`, and :user:`Erich Schubert <kno10>`.
56-
57-
:mod:`sklearn.preprocessing`
58-
............................
59-
60-
- |Efficiency| Make :class:`preprocessing.MultiLabelBinarizer` to cache class mappings instead
61-
of calculating it every time on the fly.
62-
:issue:`12116` by :user:`Ekaterina Krivich <kiote>` and `Joel Nothman`_.
63-
64-
- |Efficiency| :class:`preprocessing.PolynomialFeatures` now supports compressed
65-
sparse row (CSR) matrices as input for degrees 2 and 3. This is typically much
66-
faster than the dense case as it scales with matrix density and expansion degree
67-
(on the order of density^degree), and is much, much faster than the compressed
68-
sparse column (CSC) case. :issue:`12197` by :user:`Andrew Nystrom <awnystrom>`.
6950

70-
- |Efficiency| |API| Speed improvement in :class:`preprocessing.PolynomialFeatures`,
71-
in the dense case. Also added a new parameter ``order`` which controls output
72-
order for further speed performances. :issue:`12251` by `Tom Dupre la Tour`_.
73-
74-
:mod:`sklearn.pipeline`
75-
.......................
76-
77-
- |API| :class:`pipeline.Pipeline` now supports using ``'passthrough'`` as a
78-
transformer. :issue:`11144` by :user:`thomasjpfan`.
79-
80-
:mod:`sklearn.tree`
81-
...................
82-
- Decision Trees can now be plotted with matplotlib using
83-
:func:`tree.export.plot_tree` without relying on the ``dot`` library,
84-
removing a hard-to-install dependency.
85-
:issue:`8508` by `Andreas Müller`_.
86-
87-
- |Feature| ``get_n_leaves()`` and ``get_depth()`` have been added to
88-
:class:`tree.BaseDecisionTree` and consequently all estimators based
89-
on it, including :class:`tree.DecisionTreeClassifier`,
90-
:class:`tree.DecisionTreeRegressor`, :class:`tree.ExtraTreeClassifier`,
91-
and :class:`tree.ExtraTreeRegressor`.
92-
:issue:`12300` by :user:`Adrin Jalali <adrinjalali>`.
51+
:mod:`sklearn.linear_model`
52+
...........................
9353

54+
- |Fix| Fixed a bug in :class:`linear_model.MultiTaskElasticNet` and
55+
:class:`linear_model.MultiTaskLasso` which were breaking when
56+
``warm_start = True``. :issue:`12360` by :user:`Aakanksha Joshi <joaak>`.
9457

9558
:mod:`sklearn.metrics`
9659
......................
@@ -99,17 +62,14 @@ Support for Python 3.4 and below has been officially dropped.
9962
``'max_error'`` scorer for single output regression.
10063
:issue:`12232` by :user:`Krishna Sangeeth <whiletruelearn>`.
10164

102-
:mod:`sklearn.metrics`
103-
......................
104-
10565
- |Feature| Add :func:`metrics.multilabel_confusion_matrix`, which calculates a
10666
confusion matrix with true positive, false positive, false negative and true
10767
negative counts for each class. This facilitates the calculation of set-wise
10868
metrics such as recall, specificity, fall out and miss rate.
10969
:issue:`11179` by :user:`Shangwu Yao <ShangwuYao>` and `Joel Nothman`_.
11070

11171
:mod:`sklearn.neighbors`
112-
......................
72+
........................
11373

11474
- |API| Methods in :class:`neighbors.NearestNeighbors` :
11575
:func:`~neighbors.NearestNeighbors.kneighbors`,
@@ -120,6 +80,42 @@ Support for Python 3.4 and below has been officially dropped.
12080
when called before fit :issue:`12279` by :user:`Krishna Sangeeth
12181
<whiletruelearn>`.
12282

83+
:mod:`sklearn.pipeline`
84+
.......................
85+
86+
- |API| :class:`pipeline.Pipeline` now supports using ``'passthrough'`` as a
87+
transformer. :issue:`11144` by :user:`Thomas Fan <thomasjpfan>`.
88+
89+
:mod:`sklearn.preprocessing`
90+
............................
91+
92+
- |Efficiency| Make :class:`preprocessing.MultiLabelBinarizer` to cache class
93+
mappings instead of calculating it every time on the fly.
94+
:issue:`12116` by :user:`Ekaterina Krivich <kiote>` and `Joel Nothman`_.
95+
96+
- |Efficiency| :class:`preprocessing.PolynomialFeatures` now supports compressed
97+
sparse row (CSR) matrices as input for degrees 2 and 3. This is typically much
98+
faster than the dense case as it scales with matrix density and expansion degree
99+
(on the order of density^degree), and is much, much faster than the compressed
100+
sparse column (CSC) case. :issue:`12197` by :user:`Andrew Nystrom <awnystrom>`.
101+
102+
- |Efficiency| |API| Speed improvement in :class:`preprocessing.PolynomialFeatures`,
103+
in the dense case. Also added a new parameter ``order`` which controls output
104+
order for further speed performances. :issue:`12251` by `Tom Dupre la Tour`_.
105+
106+
:mod:`sklearn.tree`
107+
...................
108+
- Decision Trees can now be plotted with matplotlib using
109+
:func:`tree.export.plot_tree` without relying on the ``dot`` library,
110+
removing a hard-to-install dependency. :issue:`8508` by `Andreas Müller`_.
111+
112+
- |Feature| ``get_n_leaves()`` and ``get_depth()`` have been added to
113+
:class:`tree.BaseDecisionTree` and consequently all estimators based
114+
on it, including :class:`tree.DecisionTreeClassifier`,
115+
:class:`tree.DecisionTreeRegressor`, :class:`tree.ExtraTreeClassifier`,
116+
and :class:`tree.ExtraTreeRegressor`.
117+
:issue:`12300` by :user:`Adrin Jalali <adrinjalali>`.
118+
123119
Multiple modules
124120
................
125121

@@ -132,4 +128,4 @@ These changes mostly affect library developers.
132128
:func:`~utils.estimator_checks.check_estimator`, which checks that
133129
when `fit` is called twice with the same data, the ouput of
134130
`predit`, `predict_proba`, `transform`, and `decision_function` does not
135-
change. :issue:`12328` by :user:`Nicolas Hug<NicolasHug>`
131+
change. :issue:`12328` by :user:`Nicolas Hug <NicolasHug>`

0 commit comments

Comments
 (0)
0