Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

FIX several bugs in initial predictions for GBDT #12983

Merged
merged 46 commits into from Mar 1, 2019
Merged
Show file tree
Hide file tree
Changes from 33 commits
Commits
Show all changes
46 commits
Select commit Hold shift + click to select a range
ab3ac0e
fix grandient boosting with init
jeremiedbb Oct 22, 2018
2038133
add solver to avoid deprecation warning
jeremiedbb Oct 22, 2018
7ed0eb3
optional sample_weight
jeremiedbb Oct 23, 2018
65fc3e5
use has_fit_parameter instead
jeremiedbb Oct 24, 2018
ad6efdb
typo
jeremiedbb Oct 29, 2018
048dd79
try except instead of has_fit_parameter
jeremiedbb Oct 29, 2018
e0db2b2
fix merge conflicts
jeremiedbb Nov 8, 2018
a870e9c
adress comments
jeremiedbb Nov 18, 2018
5a9ce38
support multiclass classification
jeremiedbb Nov 30, 2018
bb581c0
use is_multi_class
jeremiedbb Nov 30, 2018
1129cb1
merge master
jeremiedbb Dec 17, 2018
32327ca
Merge branch 'master' into fix_gbdt_init
NicolasHug Jan 14, 2019
1efb346
Moved losses into losses.py and deprecated those in gradient_boosting.py
NicolasHug Jan 14, 2019
0171aa6
Done BinomialDeviance
NicolasHug Jan 14, 2019
a58b797
Done least squares
NicolasHug Jan 14, 2019
2df3a25
Done absolute error
NicolasHug Jan 14, 2019
b7736ef
Done Huber and Quantile losses
NicolasHug Jan 14, 2019
57f9e56
numerical stability for log
NicolasHug Jan 14, 2019
440fce9
numerical stability for log
NicolasHug Jan 14, 2019
ce03ac6
exp loss, pep8
NicolasHug Jan 14, 2019
1045796
pep8
NicolasHug Jan 15, 2019
97f68b2
properly exposed the 'zero' init estimator
NicolasHug Jan 15, 2019
bc1c652
Some doc
NicolasHug Jan 15, 2019
5b3b9fb
Fixed docstrings tests
NicolasHug Jan 15, 2019
d9b1f22
Renamed relevant things into raw_predictions
NicolasHug Jan 15, 2019
458fcb1
Updated whatsnew with multiclass classif default init bugfix
NicolasHug Jan 15, 2019
a0db3bd
Update sklearn/ensemble/losses.py
glemaitre Jan 28, 2019
3e74dae
Apply suggestions from code review
glemaitre Jan 28, 2019
e6d8bde
Addressed comments from Guillaume
NicolasHug Jan 28, 2019
2376d2c
Apply suggestions from code review
glemaitre Jan 28, 2019
501693f
Addressed comments
NicolasHug Jan 28, 2019
f210f06
Merge branch 'fix_gbdt_init' of github.com:NicolasHug/scikit-learn in…
NicolasHug Jan 28, 2019
be416e8
pep8
NicolasHug Jan 28, 2019
b875c8e
put init_raw_pred() in baseclass for regression
NicolasHug Jan 29, 2019
40bf3e3
Renamed losses.py into _gb_losses.py
NicolasHug Jan 29, 2019
dddbbd3
Merge branch 'master' into fix_gbdt_init
NicolasHug Feb 4, 2019
c70dee9
Merge branch 'master' into fix_gbdt_init
NicolasHug Feb 7, 2019
b317e27
Addressed Thomas and Jeremie comments
NicolasHug Feb 7, 2019
34619c0
Merge branch 'master' into fix_gbdt_init
NicolasHug Feb 8, 2019
2c06512
Added tests for expected output of init_raw_predictions
NicolasHug Feb 11, 2019
e2e1a5c
Added checks and tests for init estimators output shape
NicolasHug Feb 11, 2019
3078c55
Should fix FutureWarning errors
NicolasHug Feb 12, 2019
b04bcb1
Handled missing classes after early stopping differently
NicolasHug Feb 20, 2019
50310ef
Merge branch 'master' into fix_gbdt_init
NicolasHug Feb 28, 2019
d122729
EOF
NicolasHug Feb 28, 2019
f8dcfb4
Merge branch 'master' into fix_gbdt_init
jnothman Mar 1, 2019
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
12 changes: 12 additions & 0 deletions doc/whats_new/v0.21.rst
Expand Up @@ -77,10 +77,22 @@ Support for Python 3.4 and below has been officially dropped.
communication overhead. :issue:`12543` by :user:`Isaac Storch <istorch>`
and `Olivier Grisel`_.

- |Fix| Fixed a bug in :class:`ensemble.GradientBoostingClassifier` and
:class:`ensemble.GradientBoostingRegressor`, which didn't support
scikit-learn estimators as the initial estimator. Also added support of
initial estimator which does not support sample weights. :issue:`12436` by
:user:`Jérémie du Boisberranger <jeremiedbb>` and :issue:`12983` by
:user:`Nicolas Hug<NicolasHug>`.

- |Fix| Fixed a bug in :class:`ensemble.GradientBoostingClassifier` where
the gradients would be incorrectly computed in multiclass classification
problems. :issue:`12715` by :user:`Nicolas Hug<NicolasHug>`.

- |Fix| Fixed a bug in :class:`ensemble.GradientBoostingClassifier` where
the default initial prediction of a multiclass classifier would predict the
classes priors instead of the log of the priors. :issue:`12983` by
:user:`Nicolas Hug<NicolasHug>`.

- |Fix| Fixed a bug in :mod:`ensemble` where the ``predict`` method would
error for multiclass multioutput forests models if any targets were strings.
:issue:`12834` by :user:`Elizabeth Sander <elsander>`.
Expand Down