Commit Graph

77 Commits

Author SHA1 Message Date
Mathieu Blondel 0c2b07f492 Fix warning message. 2012-09-29 02:16:06 +09:00
Mathieu Blondel 2a7ebd955f Cosmit: not a fan of defining a function in a loop :) 2012-09-29 02:16:06 +09:00
Mathieu Blondel 7575410adc PEP8! 2012-09-29 02:16:06 +09:00
Mathieu Blondel 71fd662f63 Better documentation on the choice of solver. 2012-09-29 02:16:06 +09:00
Mathieu Blondel 15ce0557b2 lsqr is not available in old scipy versions... 2012-09-29 02:16:06 +09:00
Mathieu Blondel 1e117305b6 Fall back to dense_cholesky if sample_weight is given. 2012-09-29 02:16:05 +09:00
Mathieu Blondel c5892aabb8 Fix damp parameter. 2012-09-29 02:16:05 +09:00
Mathieu Blondel abee624bf6 Fix docstrings and doctests. 2012-09-29 02:16:05 +09:00
Mathieu Blondel 047109233f BUG: transmit parameters correctly from Ridge to ridge_regression. 2012-09-29 02:16:05 +09:00
Mathieu Blondel 107abb6f3e Add lsqr solver. 2012-09-29 02:16:05 +09:00
Mathieu Blondel 2cae34b059 Move solver option to constructor. 2012-09-29 02:16:05 +09:00
Lars Buitinck d93b33ed1a ENH add max_iter argument to Ridge estimators 2012-09-19 14:08:52 +02:00
Fabian Pedregosa 33efa17665 ENH Improve Ridge's conjugate gradient descent
Reorder terms in the algorithm so that X'X is not explicitly computed and
add LinearOperator as accepted types. This improves precedent code in two ways:

    * Speed: I get speedups ranging from 2.5x to more than 10x.
      Benchmark suite can be seen here: https://gist.github.com/1322711
      where I print the time spent for various sizes and densities.

    * Interface. By allowing a LinearOperator type, we can use the
      conjugate gradient method almost for free on dense matrices with
      special structure, pytables, etc., about anything that
      implements a fast matrix-vector product.
2012-09-19 14:08:52 +02:00
Lars Buitinck 35704c8351 COSMIT rm deprecated class_weight from fit in Ridge 2012-09-19 14:08:41 +02:00
Lars Buitinck 4183cb3b34 COSMIT use LinearClassifierMixin in RidgeClassifier
Had to change a test due to different tie breaking.
2012-09-17 23:22:16 +02:00
Andreas Mueller 713ba6006e COSMIT pep8 2012-09-04 17:24:01 +01:00
Andreas Mueller 63e99023b6 DOC fixed some sphinx errors, issues in docs.... 2012-09-03 23:48:32 +01:00
Alexandre Gramfort 466479fe03 API : consistent use alpha_/alphas_ for alpha/alphas estimated by CV in linear models (issue #1041) 2012-09-02 23:16:54 +02:00
Vlad Niculae 9e7c03f71a Add or fix deprecation schedule in warnings. 2012-09-01 20:02:14 +02:00
Andreas Mueller 0fccfbe95c COSMIT trying to fix doc issues 2012-09-01 16:23:57 +02:00
Mathieu Blondel 920bf84e67 Decision threshold is now 0 in RidgeClassifier. 2012-08-30 22:42:56 +09:00
Mathieu Blondel e18a32d93b Fix #1080. 2012-08-28 21:33:35 +09:00
Nicolas Pinto 129fd7b7bf COSMIT: -loo_values +cv_values 2012-07-18 07:42:04 +02:00
Nicolas Pinto cfb6510729 COSMIT: rename M matrix to loo_values 2012-07-18 07:42:04 +02:00
Nicolas Pinto 29d5dc2f09 FIX: expose loo_values_ in RidgeCV instead of the private _RidgeGCV 2012-07-18 07:42:04 +02:00
Nicolas Pinto 602725274f ENH: add store_loo_values attribute to _RidgeGCV see Issue #957 2012-07-18 07:42:04 +02:00
Charles-Pierre Astolfi b07dfafbf0 Typo fix 2012-07-11 01:47:48 +02:00
Andreas Mueller 94708a33b9 FIX minor rebasing and other problems 2012-06-26 19:00:51 +02:00
Andreas Mueller 971d1310b9 ENH create BaseRidge base class 2012-06-26 14:16:04 +02:00
Mathieu Blondel 7feff8ceb2 Fix: RidgeClassifier must not inherit from RegressorMixin. 2012-06-14 14:05:09 +09:00
Mathieu Blondel 5709d53134 Removed needless method redefinition. 2012-06-13 15:28:38 +09:00
Mathieu Blondel fb714df433 Fix #904. 2012-06-13 13:26:02 +09:00
Andreas Mueller 705a36801b DOC Believe it or not - this fixes the annoying sphinx error. And don't dare to
abbreviate singular value decomposition with svd. Sphinx does not approve of that.
2012-03-31 16:23:01 +02:00
Emmanuelle Gouillart e9d6673c6e ENH: a few typos in docstrings 2012-03-25 23:17:42 +02:00
Andreas Mueller 57e81471a5 Merge pull request #648 from amueller/linear_model_class_weights
MRG class_weights constructor parameter in RidgeCV
2012-03-25 07:54:45 -07:00
Andreas Mueller fb4c81bb49 ENH local variable in ``fit`` instead of modifying the estimator parameters. thanks @GaelVaroquaux 2012-03-21 23:45:02 +01:00
Andreas Mueller 667ef37714 DOC move references from Notes to References section in docstrings 2012-03-04 15:16:51 +01:00
Andreas Mueller 689d672529 ENH added "shadow" parameter class_weight_ as @ogrisel suggested. 2012-03-04 12:18:34 +01:00
Andreas Mueller 9061131d7f Merge branch 'master' into linear_model_class_weights
Conflicts:
	sklearn/linear_model/ridge.py
2012-03-03 17:01:50 +01:00
Andreas Mueller 4508438445 ENH Added tests and fixes 2012-03-03 15:02:48 +01:00
Andreas Mueller c597c6b5ae COSMIT pep8 2012-03-03 14:40:51 +01:00
Andreas Mueller 5886f7cacc DOC add removed docstring. 2012-03-03 14:39:17 +01:00
Andreas Mueller 52263a60f4 DOC Documentation for RidgeClassifierCV 2012-03-03 14:00:56 +01:00
Andreas Mueller 6bf9501a9a FIX: class_weight only in classifier Ridge classes 2012-03-03 13:54:10 +01:00
Andreas Mueller bc942e8c9d COSMIT pep8 2012-03-02 19:39:43 +01:00
Andreas Mueller 6783466bb0 DOC doc fixes 2012-02-29 22:18:18 +01:00
Mathieu Blondel 36602c6cb4 Don't raise warning in auto mode. 2012-02-28 20:46:18 +09:00
Mathieu Blondel cbb67aa7d3 Pep8. 2012-02-28 20:46:18 +09:00
fcostin 8195db25af refactoring (thanks @GaelVaroquaux, @mblondel) 2012-02-28 20:46:18 +09:00
fcostin 83ecb9bebc coerce sparse matrices to dense before SVD
...all ridge tests pass now.
2012-02-28 20:46:18 +09:00