Alexis Mignon
07bd149b68
Created decorator function so that custom objective function passed to the constructor are more consistent with the sklearn conventions. Added comments in the doc string
2016-02-16 10:58:22 +01:00
Alexis Mignon
c8714f587a
Added the possibility to use custom objective function in the sklearn API
2016-02-15 17:13:13 +01:00
Maxim Grechkin
f5e96eba72
Make missing handling consistent with sklearn's portion of the python package
2016-01-28 14:16:11 -08:00
Kai Luo
d9e50fd7f3
__copy__ calls __deepcopy__ with an argument
2016-01-20 19:57:20 +08:00
Kai Luo
5cd765e935
fix signature of __deepcopy__ method
2016-01-20 17:18:11 +08:00
tqchen
8e7f2679d5
[DOC] Update R doc
2016-01-16 11:52:33 -08:00
tqchen
1495a43cea
[R] make all customizations to meet strict standard of cran
2016-01-16 10:25:12 -08:00
tqchen
d75e3ed05d
[LIBXGBOOST] pass demo running.
2016-01-16 10:24:01 -08:00
Yuan (Terry) Tang
7606bf8156
Fixes #725
2016-01-06 18:21:29 -06:00
FrozenFingerz
2a46918c66
python cv bugfixing
...
- fixed bug if both eval_metrics xgb-param and
metrics param of cv function have been set
- cv early stopping output looks now like the one of xgb.train
2015-12-29 12:24:38 +01:00
FrozenFingerz
38b773d80b
cv: fixed devision by zero exception
...
- show_progress=False or show_progress=0 led to devision by zero exception
2015-12-27 13:54:52 +01:00
Faron
b3f3e7d0cb
fixed wrong iter when using training continuation
2015-12-19 10:35:16 +01:00
Randy Carnevale
380e54a753
docstring typo
2015-12-16 17:25:55 -05:00
Randy Carnevale
0825ab36f0
updating docs for cv
2015-12-16 17:21:23 -05:00
Randy Carnevale
a3fe14d6c6
modifying cv show_progress to allow print-every-n behavior
2015-12-16 16:33:01 -05:00
Ajinkya Kale
0772b51c2c
minor change dir
2015-12-12 16:34:07 -08:00
Ajinkya Kale
4695fa3c2a
adding right path to setup.py
2015-12-12 15:08:59 -08:00
terrytangyuan
0eb6240fd0
Fixed all lint errors
2015-12-11 18:46:15 -06:00
terrytangyuan
a7e79e089b
fix lint errors in core
2015-12-11 18:37:13 -06:00
terrytangyuan
7be496a051
ignore nested blocks
2015-12-11 18:20:35 -06:00
terrytangyuan
c3ec8ee76f
Added pylintrc file
2015-12-11 18:10:15 -06:00
phunterlau
a4840b0268
update pip building, troubleshooting with new makefile, plus friendly error message when fail importing sklearn
2015-12-07 22:29:46 -08:00
Yuan (Terry) Tang
162e91c5ca
change .md to .rst
2015-12-06 20:25:53 -06:00
Yuan (Terry) Tang
e25b2c4968
Remove redundant README
2015-12-06 11:05:44 -05:00
Far0n
af166bf0a0
small verbose_eval fixes
...
- ensures same behavior for verbose_eval=0 and verbose_eval=False
- fix printing last eval message if early_stopping_rounds is set, but xgb
runs to the end
2015-11-24 09:22:25 +01:00
Johan Manders
e68e9659ab
Python verbose_eval extension
...
This is an extension of the verbose_eval abilities.
Removed some trailing-whitespaces
2015-11-13 08:19:44 +01:00
sinhrks
25c4fbd0cb
Cleanup pandas support
2015-11-13 06:55:04 +09:00
Yuan (Terry) Tang
4fb6153eed
Fixed minor lint issue
2015-11-12 09:01:05 -06:00
Yuan (Terry) Tang
0a0951ba12
Clarification for best_ntree_limit
2015-11-12 08:53:45 -06:00
Yuan (Terry) Tang
42e1fd8fff
Merge pull request #598 from Far0n/py_train
...
best_ntree_limit attribute & training continuation bugfix
2015-11-12 06:16:19 -06:00
phunterlau
ee4096d23e
fix pushd problem of pip building, convert README to rst for PyPI
2015-11-11 23:03:07 -08:00
Yuan (Terry) Tang
7b3fd92015
Added PyPI badges
2015-11-10 18:23:39 -06:00
Far0n
ce5930c365
best_ntree_limit attribute added
...
- best_ntree_limit as new booster atrribute added
- usage of bst.best_ntree_limit in python doc added
- fixed wrong 'best_iteration' after training continuation
2015-11-10 15:37:22 +01:00
antonymayi
8c7b18daed
python 2.6 compatibility tweak
...
replacing set literal {} with set() for python 2.6 compatibility (plus reformatting the line)
2015-11-10 14:50:54 +01:00
Yuan (Terry) Tang
1dd96b6cdc
Merge pull request #597 from JohanManders/python-pandas-dtypes
...
Python pandas dtypes
2015-11-09 18:08:41 -06:00
antonymayi
7114d6681a
Update training.py
...
pylint compliancy
2015-11-09 15:09:14 +01:00
antonymayi
34e01642ca
Update training.py
...
avoid dict comprehension for python 2.6 compatibility
2015-11-09 14:26:16 +01:00
Yuan (Terry) Tang
b8bc85b534
Clarification for learning_rates
2015-11-08 21:10:04 -06:00
Faron
b2f98db74e
grammar correction
2015-11-08 21:00:16 +01:00
Yuan (Terry) Tang
01053f8f2f
Merge pull request #594 from Far0n/feval
...
python: multiple eval_metrics changes
2015-11-08 10:10:28 -05:00
FrozenFingerz
b59018aa05
python: multiple eval_metrics changes
...
- allows feval to return a list of tuples (name, error/score value)
- changed behavior for multiple eval_metrics in conjunction with
early_stopping: Instead of raising an error, the last passed evel_metric
(or last entry in return value of feval) is used for early stopping
- allows list of eval_metrics in dict-typed params
- unittest for new features / behavior
documentation updated
- example for assigning a list to 'eval_metric'
- note about early stopping on last passed eval metric
- info msg for used eval metric added
2015-11-08 11:23:54 +01:00
Faron
95cc900b1f
early stopping for CV (python)
2015-11-07 09:52:36 +01:00
Johan Manders
5f0f8749d9
Cleaned up some code
2015-11-04 18:05:47 +01:00
Johan Manders
f9e1b2b7b7
Added back feature names
2015-11-03 21:26:11 +01:00
Johan Manders
96f221e0d0
Merge pull request #5 from dmlc/master
...
Update to latest version
2015-11-03 20:37:20 +01:00
Far0n
b894f7c9d6
bugfix type-check xgb_model param
2015-11-03 14:43:08 +01:00
Faron
4fe2f2fb09
python train additions
...
+ training continuation of existing model
+ maximize parameter just like in R package (whether to maximize feval)
2015-11-02 21:21:05 +01:00
Faron
79813097b5
sklearn_wrapper additions
...
added output_margin & ntree_limit to predict and predict_proba
2015-11-02 17:41:30 +01:00
phunterlau
739b3f2c5f
separate setup.py with pip installation, add trouble shooting page
2015-11-01 22:11:11 -08:00
terrytangyuan
610b70b79e
Suppress more evaluation verbose during training
2015-10-31 13:05:52 -04:00
