9ad829b307
Merge pull request #1914 from fabianp/ridge_svd
...
MRG - Add SVD-based solver to ridge regression.
2013-06-06 03:12:54 -07:00
9589e04fd9
FIX typo in test for RdigeCV
2013-05-07 22:52:41 +02:00
daf527735a
BUG disable memory-blowing SVD for sparse input in RidgeCV
...
Fixes #1921 .
2013-05-03 17:48:39 +02:00
922df5ae5e
Add SVD-based solver to ridge regression.
...
I added a solver based on the thin SVD of X to compute the ridge
coefficients. This is much more stable than the cholesky decomposition
for singular matrices.
While in the Ridge regression the Gram matrix becomes nonsingular
because of the regularization, I've come across problems when the
regularization parameter is very small. This pull request remedies,
making the algorithm defined and stable even for the border case
alpha=0.
I took the liberty to change the default algorithm of to this one
(when there are no sample weights).
2013-05-02 11:30:18 +02:00
dacfd8bd5d
DOC: Replaced all BSD style licenses with "BSD 3 clause"
...
Replaced all BSD style licenses with "BSD 3 clause"
Not checked yet!
Removed duplicate "3 clause, 3 clause"
Removed trailing period if exists
Fixed some missed licences, still about 50 to do, but those can be automated
Think I got the last of them.
Apparently me and sed have different ideas of regex.
Found a few more
2013-04-30 08:34:46 +02:00
465e053c49
COSMIT pep8
2013-04-28 15:46:13 +02:00
67ece78da4
COSMIT various typofixes
...
As suggested by codespell <https://github.com/lucasdemarchi/codespell >
Signed-off-by: Stefano Lattarini <stefano.lattarini@gmail.com>
2013-04-11 21:22:14 +02:00
972e7cffa4
BUG: ElasticNectCV choosing improper l1_ratio
...
The code was lacking good tests: it had only smoke tests. Shame on
me (I am the author).
2013-04-03 09:48:28 +02:00
a1d4d184fe
COSMIT pep8
2013-04-02 13:30:55 +02:00
01801886b2
ENH multiclass probability estimates for SGDClassifier
...
Fixes #1814 .
2013-03-30 12:33:33 +01:00
bf0c1a38bb
Merge branch 'master' into pr/1798
2013-03-25 08:41:43 +01:00
154007b0ac
fixing bug in linear_model.SGDClassifier for multi-class warm start
2013-03-21 03:03:38 -07:00
5a5268b89f
PY3: fix remaining cStringIO imports
2013-03-19 17:57:12 +01:00
cdb9ef2d58
COSMIT prevent a copy in randomized LR
...
Also remove some superseded Py3 compat code.
2013-03-17 12:30:17 +01:00
f95dfe7763
BUG SGDClassifier and friends did not forget labels_ in re-fit
...
Fixes #1703 .
2013-03-15 12:37:16 +01:00
b8d23fa85a
FIX : test due to KFold change
2013-03-02 16:39:42 +01:00
2afe4d5a63
TST/FIX always use "almost equal" for floats.
2013-03-01 12:14:50 +01:00
b83f1120e1
TST/FIX set random_state in logistic regression tests
2013-03-01 11:59:03 +01:00
6bf4c90c5d
ENH sparsify method for L1-reg linear models
...
Currently implemented on LinearSVC, LR and SGD.
2013-02-23 12:11:56 +01:00
541987802f
P3K range vs. xrange
2013-02-14 02:05:35 +01:00
85ec0fd1ae
P3K death to the print statement
2013-02-11 23:11:57 +01:00
1967a0b323
P3K use six to have a python 2 & 3 compatible code base
2013-02-09 18:08:44 +01:00
bdf2d4b85a
TST some more tests for multi output lars
2013-01-29 23:52:55 +01:00
e52e24706f
COSMIT pep8
2013-01-21 21:20:21 +01:00
e3dd35fa11
BUG: shape bug in tests
2013-01-20 19:16:18 +01:00
94404fddea
TST: improve last test
2013-01-20 19:16:18 +01:00
c050d28d94
TEST: add a test of sample weights
2013-01-20 19:16:18 +01:00
906290e57e
ENH: rng local to tests
...
Elsewhere individual tests may not be reproducible
2013-01-20 19:16:18 +01:00
fd117f1e51
TST catch expected warning in sparse coordinate descent test.
2013-01-17 20:23:40 +01:00
966e10c252
COSMIT PEP8
2012-12-22 16:48:17 +01:00
754c2d9533
COSMIT pep8
2012-12-22 16:48:17 +01:00
270bc8e278
TST some more tests for SGDClassifier input validation
2012-12-15 22:10:48 +01:00
8eda5e1c0e
TST document and test verbosity parameter of lars_path
2012-12-15 21:18:10 +01:00
08dff76e2c
ENH input validation only in ``fit`` in LassoLarsIC, check that error is raised.
2012-12-15 21:17:51 +01:00
116b54a8c8
Fix bug in sqnorm (used by PassiveAggressive).
...
The function was using the wrong data indices T_T.
2012-11-24 23:05:52 +09:00
23ac37a4fb
API : move isotonic regression out of linear_model
2012-11-14 22:01:53 +01:00
c97e281401
FIX #1354 : machine precision assertion failure in test_liblinear_random_state
2012-11-10 20:54:07 +01:00
c9cd2b7292
FIX: lars drop for good platform specific test failure
2012-11-10 20:41:21 +01:00
7cec21bbee
TEST: decrease precision in test_lars_drop_for_good
...
If the test is wrong, this will be well off. Small errors just reflect
numerical errors
2012-11-10 19:43:17 +01:00
92138a4258
COSMIT pep8
2012-11-10 14:02:29 +00:00
9e52350533
BUG: compat with numpy 1.3
2012-11-08 11:02:54 +01:00
25707c2c54
BUG: missing import introduced by rebase
2012-11-08 10:40:10 +01:00
de4fd325af
COSMIT: better comments
...
Specifying exactly why a test is around is important
2012-11-08 10:39:22 +01:00
1f3f49755a
TST: difficult test for early stopping
2012-11-08 10:39:22 +01:00
428439f7ca
MISC: comment
2012-11-08 10:39:22 +01:00
2d9435c366
TST: add a test for lasso and lars
2012-11-08 10:39:22 +01:00
e290951871
ENH: early stopping for lars
...
When the noise in the updates of the Cholesky becomes greater than the
max residual, it is impossible to go on in the path: giving up.
2012-11-08 10:39:22 +01:00
f739070d5f
WIP: drop for good correlated regressors
...
Attempt to drop for good correlated regressors. In Lars. Doesn't seem to
be working, as it is give a significantly different result than the
lasso (failing test)
2012-11-08 10:39:22 +01:00
a911ad10fb
MISC: address comments
2012-11-08 10:32:58 +01:00
156566e6a9
ENH: early stopping LARS for degenerate active set
2012-11-08 10:31:13 +01:00
Fabian Pedregosa