Andreas Mueller
|
676b97fad6
|
FIX rebase mishap
|
2012-10-23 14:13:08 +02:00 |
|
Jacques Kvam
|
29540793c4
|
fix GradientBoostingClassifier by passing verbose as a keyword argument
|
2012-10-23 14:13:08 +02:00 |
|
Jacques Kvam
|
513d2d91c1
|
remove '\r' and format numbers to be fixed width, 7 digits of precision
|
2012-10-23 14:13:08 +02:00 |
|
Jacques Kvam
|
5dfcff0549
|
Changed verbose to int, added a low verbose option to just print '.'.
|
2012-10-23 14:13:08 +02:00 |
|
Jacques Kvam
|
67890e4047
|
add verbose output for gradient boosting algorithms
|
2012-10-23 14:13:08 +02:00 |
|
Peter Prettenhofer
|
b321d4c893
|
added Emanuele to authors
|
2012-10-18 09:40:39 +02:00 |
|
Peter Prettenhofer
|
ff3097f236
|
cosmit
|
2012-10-18 09:30:18 +02:00 |
|
Peter Prettenhofer
|
bfbcc787b9
|
docstrings for init and loss_
|
2012-10-18 09:06:56 +02:00 |
|
Peter Prettenhofer
|
6ac7fbfc97
|
Merge branch 'master' into pr/975
Conflicts:
sklearn/ensemble/gradient_boosting.py
|
2012-10-18 08:53:24 +02:00 |
|
Andreas Mueller
|
713ba6006e
|
COSMIT pep8
|
2012-09-04 17:24:01 +01:00 |
|
Peter Prettenhofer
|
edee5cacc9
|
rm CV class legacy
|
2012-09-04 07:56:53 +02:00 |
|
Peter Prettenhofer
|
e6f0b1d664
|
remove *CV classes - only pick decision_function and staged predict refactoring
|
2012-09-04 07:56:52 +02:00 |
|
Peter Prettenhofer
|
970326fe12
|
wip: fixing some set attr errors but still buggy if params not lists
|
2012-09-04 07:56:52 +02:00 |
|
Peter Prettenhofer
|
e93a6a509f
|
wip: aggregate fold via groupby
|
2012-09-04 07:56:52 +02:00 |
|
Peter Prettenhofer
|
d7fc786273
|
GradientBoosting classes with built in cross-validation; implemented via Decorator pattern
|
2012-09-04 07:56:52 +02:00 |
|
Peter Prettenhofer
|
554799fcc5
|
ENH: use gini for feature importance
|
2012-09-04 07:56:52 +02:00 |
|
Peter Prettenhofer
|
425d1f9a2e
|
refactored prediction and decision_function (rm duplicate code)
|
2012-09-04 07:56:52 +02:00 |
|
Peter Prettenhofer
|
bd4292a50c
|
work on BaseGradientBoostingCV
|
2012-09-04 07:56:52 +02:00 |
|
Gael Varoquaux
|
a12bd7bc7c
|
ENH: understandable error message for X sparse
|
2012-08-10 11:23:00 +01:00 |
|
emanuele
|
8db20651be
|
FIX: added logsumexp and nan_to_num to avoid underflows and NaNs
|
2012-07-23 17:19:20 +02:00 |
|
Peter Prettenhofer
|
2cce14c6ef
|
cosmit: better docstrings
|
2012-07-21 16:01:17 +02:00 |
|
Gilles Louppe
|
2fe48dc0fd
|
Tree refactoring (19)
|
2012-07-16 13:47:20 +02:00 |
|
Gilles Louppe
|
1f282c6186
|
FIX: gradient boosting (1)
|
2012-07-16 13:07:18 +02:00 |
|
Gilles Louppe
|
532c54c5ac
|
Merge branch 'master' of github.com:scikit-learn/scikit-learn into tree-mo
Conflicts:
sklearn/tree/_tree.c
|
2012-07-09 17:32:43 +02:00 |
|
Peter Prettenhofer
|
545f324a2d
|
y must be one dimensional
|
2012-07-05 22:07:22 +02:00 |
|
Gilles Louppe
|
41cd38f8bc
|
ENH: Patch GradientBoosting (2)
|
2012-07-02 11:37:28 +02:00 |
|
Gilles Louppe
|
5ee718c308
|
ENH: Patch GradientBoosting
|
2012-07-02 11:31:09 +02:00 |
|
Peter Prettenhofer
|
8792cbd6f7
|
Merge remote branch 'upstream/master' into gbrt-huber
|
2012-07-02 08:58:50 +02:00 |
|
Andreas Mueller
|
26c31bf810
|
ENH make base classes abstract meta classes
|
2012-06-26 14:16:03 +02:00 |
|
Peter Prettenhofer
|
688b612abc
|
Use BaseEstimator for constant predictors
|
2012-06-14 21:28:11 +02:00 |
|
Peter Prettenhofer
|
505abb448b
|
import random split
|
2012-06-06 13:41:08 +02:00 |
|
Peter Prettenhofer
|
aa8354a4ab
|
wip: added quantile regression loss; this allows for prediction intervals; adopted the GP regression example to show-case prediction intervals
|
2012-06-05 23:07:56 +02:00 |
|
Peter Prettenhofer
|
b6b564dc3f
|
fix: alpha needs to be scaled by 100
|
2012-06-03 11:45:33 +02:00 |
|
Peter Prettenhofer
|
48a08d76c7
|
fix: forgot comma
|
2012-06-02 17:05:36 +02:00 |
|
Peter Prettenhofer
|
fab9857249
|
added feature subsampling to GBRT (via max_features)
|
2012-06-02 17:04:26 +02:00 |
|
Peter Prettenhofer
|
196fd69809
|
Merge branch 'master' into gbrt-huber
Conflicts:
sklearn/ensemble/gradient_boosting.py
|
2012-05-30 07:54:48 +02:00 |
|
Peter Prettenhofer
|
7d89ab0949
|
fix: make sure X is c-continuous
|
2012-05-30 07:53:41 +02:00 |
|
Peter Prettenhofer
|
038a48933e
|
fix: ensure X is c-continuous
|
2012-05-30 07:48:06 +02:00 |
|
Peter Prettenhofer
|
7f436ab518
|
add alpha parameter for huber robust regression loss
|
2012-05-29 13:16:15 +02:00 |
|
Peter Prettenhofer
|
583ce7025d
|
fix errors in huber loss
|
2012-05-28 15:45:14 +02:00 |
|
Peter Prettenhofer
|
a170d84412
|
implemented huber loss for robust regression
|
2012-05-25 23:08:07 +02:00 |
|
Peter Prettenhofer
|
bc9f7992f0
|
started work on Huber loss function for robust regression
|
2012-05-23 10:00:24 +02:00 |
|
Peter Prettenhofer
|
09d424166f
|
fix: gradient boosting regressor does not check if X is c-continous
|
2012-05-12 18:29:06 +02:00 |
|
Andreas Mueller
|
fb7b5b7e76
|
DOC correct links to face recognition example, take care of trailing underscores.
|
2012-04-08 15:51:18 +02:00 |
|
Peter Prettenhofer
|
b35aa6f134
|
cosmit
|
2012-04-02 12:14:44 +02:00 |
|
Peter Prettenhofer
|
b82c76e4eb
|
better input checks at prediciton time
|
2012-03-30 08:28:15 +02:00 |
|
Peter Prettenhofer
|
54cf09f9ee
|
staged_predict for regression
|
2012-03-29 22:08:06 +02:00 |
|
Peter Prettenhofer
|
11aa1077f6
|
add staged predict
|
2012-03-29 10:26:30 +02:00 |
|
Peter Prettenhofer
|
5467daca60
|
faster prediction methods
|
2012-03-29 08:35:37 +02:00 |
|
Peter Prettenhofer
|
1bd5e8342c
|
cosmit; mention scalability issues w.r.t. large number of classes
|
2012-03-26 08:55:50 +02:00 |