8000 DOC Add what's new for 0.19.1 (#9983) · scikit-learn/scikit-learn@08dc3ff · GitHub
[go: up one dir, main page]
Skip to content

Commit 08dc3ff

Browse files
jnothmanjnothman
authored
DOC Add what's new for 0.19.1 (#9983)
1 parent 0e19710 commit 08dc3ff

File tree

1 file changed

+130
-0
lines changed

1 file changed

+130
-0
lines changed

‎doc/whats_new/v0.19.rst

Lines changed: 130 additions & 0 deletions
< 8000 td data-grid-cell-id="diff-9b0a240dee95eff0ebe4359e4572c27b03f55c5cc6e70aaf7ce3007f9b0c76d6-6-77-0" data-selected="false" role="gridcell" style="background-color:var(--diffBlob-additionNum-bgColor, var(--diffBlob-addition-bgColor-num));text-align:center" tabindex="-1" valign="top" class="focusable-grid-cell diff-line-number position-relative left-side">
Original file line numberDiff line numberDiff line change
@@ -4,6 +4,136 @@
44

55
.. _changes_0_19:
66

7+
Version 0.19.1
8+
==============
9+
10+
**October 23, 2017**
11+
12+
This is a bug-fix release with some minor documentation improvements and
13+
enhancements to features released in 0.19.0.
14+
15+
Note there may be minor differences in TSNE output in this release (due to
16+
:issue:`9623`), in the case where multiple samples have equal distance to some
17+
sample.
18+
19+
Changelog
20+
---------
21+
22+
API changes
23+
...........
24+
25+
- Reverted the addition of ``metrics.ndcg_score`` and ``metrics.dcg_score``
26+
which had been merged into version 0.19.0 by error. The implementations
27+
were broken and undocumented.
28+
29+
- ``return_train_score`` which was added to
30+
:class:`model_selection.GridSearchCV`,
31+
:class:`model_selection.RandomizedSearchCV` and
32+
:func:`model_selection.cross_validate` in version 0.19.0 will be changing its
33+
default value from True to False in version 0.21. We found that calculating
34+
training score could have a great effect on cross validation runtime in some
35+
cases. Users should explicitly set ``return_train_score`` to False if
36+
prediction or scoring functions are slow, resulting in a deleterious effect
37+
on CV runtime, or to True if they wish to use the calculated scores.
38+
:issue:`9677` by :user:`Kumar Ashutosh <thechargedneutron>` and `Joel
39+
Nothman`_.
40+
41+
- ``correlation_models`` and ``regression_models`` from the legacy gaussian
42+
processes implementation have been belatedly deprecated. :issue:`9717` by
43+
:user:`Kumar Ashutosh <thechargedneutron>`.
44+
45+
Bug fixes
46+
.........
47+
48+
- Avoid integer overflows in :func:`metrics.matthews_corrcoef`.
49+
:issue:`9693` by :user:`Sam Steingold <sam-s>`.
50+
51+
- Fix ValueError in :class:`preprocessing.LabelEncoder` when using
52+
``inverse_transform`` on unseen labels. :issue:`9816` by :user:`Charlie Newey
53+
<newey01c>`.
54+
55+
- Fixed a bug in the objective function for :class:`manifold.TSNE` (both exact
56+
and with the Barnes-Hut approximation) when ``n_components >= 3``.
57+
:issue:`9711` by :user:`goncalo-rodrigues`.
58+
59+
- Fix regression in :func:`model_selection.cross_val_predict` where it
60+
raised an error with ``method='predict_proba'`` for some probabilistic
61+
classifiers. :issue:`9641` by :user:`James Bourbeau <jrbourbeau>`.
62+
63+
- Fixed a bug where :func:`datasets.make_classification` modified its input
64+
``weights``. :issue:`9865` by :user:`Sachin Kelkar <s4chin>`.
65+
66+
- :class:`model_selection.StratifiedShuffleSplit` now works with multioutput
67+
multiclass or multilabel data with more than 1000 columns. :issue:`9922` by
68+
:user:`Charlie Brummitt <crbrummitt>`.
69+
70+
- Fixed a bug with nested and conditional parameter setting, e.g. setting a
71+
pipeline step and its parameter at the same time. :issue:`9945` by `Andreas
72+
Müller`_ and `Joel Nothman`_.
73+
74+
Regressions in 0.19.0 fixed in 0.19.1:
75+
76+
- Fixed a bug where parallelised prediction in random forests was not
77+
thread-safe and could (rarely) result in arbitrary errors. :issue:`9830` by
78+
`Joel Nothman`_.
79+
80+
- Fix regression in :func:`model_selection.cross_val_predict` where it no
81+
longer accepted ``X`` as a list. :issue:`9600` by :user:`Rasul Kerimov
82+
<CoderINusE>`.
83+
84+
- Fixed handling of :func:`cross_val_predict` for binary classification with
85+
``method='decision_function'``. :issue:`9593` by :user:`Reiichiro Nakano
86+
<reiinakano>` and core devs.
87+
88+
- Fix regression in :class:`pipeline.Pipeline` where it no longer accepted
89+
``steps`` as a tuple. :issue:`9604` by :user:`Joris Van den Bossche
90+
<jorisvandenbossche>`.
91+
92+
- Fix bug where ``n_iter`` was not properly deprecated, leaving ``n_iter``
93+
unavailable for interim use in
94+
:class:`linear_model.SGDClassifier`, :class:`linear_model.SGDRegressor`,
95+
:class:`linear_model.PassiveAggressiveClassifier`,
96+
:class:`linear_model.PassiveAggressiveRegressor` and
97+
:class:`linear_model.Perceptron`. :issue:`9558` by `Andreas Müller`_.
98+
99+
- Dataset fetchers make sure temporary files are closed before removing them,
100+
which caused errors on Windows. :issue:`9847` by :user:`Joan Massich <massich>`.
101+
102+
- Fixed a regression in :class:`manifold.TSNE` where it no longer supported
103+
metrics other than 'euclidean' and 'precomputed'. :issue:`9623` by :user:`Oli
104+
Blum <oliblum90>`.
105+
106+
Enhancements
107+
............
108+
109+
- Our test suite and :func:`utils.estimator_checks.check_estimators` can now be
110+
run without Nose installed. :issue:`9697` by :user:`Joan Massich <massich>`.
111+
112+
- To improve usability of version 0.19's :class:`pipeline.Pipeline`
113+
caching, ``memory`` now allows ``joblib.Memory`` instances.
114+
This make use of the new :func:`utils.validation.check_memory` helper.
115+
issue:`9584` by :user:`Kumar Ashutosh <thechargedneutron>`
116+
117+
- Some fixes to examples: :issue:`9750`, :issue:`9788`, :issue:`9815`
118+
119+
- Made a FutureWarning in SGD-based estimators less verbose. :issue:`9802` by
120+
:user:`Vrishank Bhardwaj <vrishank97>`.
121+
122+
Code and Documentation Contributors
123+
-----------------------------------
124+
125+
With thanks to:
126+
127+
Joel Nothman, Loic Esteve, Andreas Mueller, Kumar Ashutosh,
128+
Vrishank Bhardwaj, Hanmin Qin, Rasul Kerimov, James Bourbeau,
129+
Nagarjuna Kumar, Nathaniel Saul, Olivier Grisel, Roman
130+
Yurchak, Reiichiro Nakano, Sachin Kelkar, Sam Steingold,
131+
Yaroslav Halchenko, diegodlh, felix, goncalo-rodrigues,
132+
jkleint, oliblum90, pasbi, Anthony Gitter, Ben Lawson, Charlie
133+
Brummitt, Didi Bar-Zev, Gael Varoquaux, Joan Massich, Joris
134+
Van den Bossche, nielsenmarkus11
135+
136+
7137
Version 0.19
8138
============
9139

0 commit comments

Comments
 (0)
0