Andreas Mueller
|
896f9d97cf
|
Fix handling of Nystroem params with a callable kernel (#9166)
|
2017-06-30 18:16:10 +02:00 |
|
Romain Brault
|
15b7d47a2f
|
relax skewness assumption (#7573)
relax skewedness assumption
relax skewedness assumption
relax skewedness assumption
relax skewedness assumption
relax skewedness assumption
last corrections
whats_new
merge
complying whats_new
removed unnecessary _assert_X
increased coverage
relax skewness assumption
relax skewedness assumption
relax skewedness assumption
relax skewedness assumption
relax skewedness assumption
relax skewedness assumption
last corrections
whats_new
merge
complying whats_new
removed unnecessary _assert_X
increased coverage
remove cythonize.dat and merge
simplify tests
simplify tests
|
2017-05-18 14:17:10 -04:00 |
|
seales
|
0485ada58b
|
General spelling fixes
|
2015-12-16 09:46:42 -08:00 |
|
Raghav R V
|
cd2ee7e454
|
MAINT docstring --> comments to prevent nose from using doc in verbose mode
|
2015-03-21 11:16:49 +05:30 |
|
Andreas Mueller
|
8db11a8c74
|
Make nystroem approximation robust to singular kernel.
|
2015-01-29 12:11:52 -05:00 |
|
Dougal J. Sutherland
|
f0f26bce33
|
fix factor of 2 in RBFSampler; make test more rigorous
http://nbviewer.ipython.org/gist/dougalsutherland/11fd9895924c3c1aed19
demonstrates the problem in pictures.
|
2014-09-22 18:19:40 -04:00 |
|
Sebastian Saeger
|
5ebba54259
|
Fix the pep8 violations
|
2014-04-17 20:50:07 +02:00 |
|
Sebastian Saeger
|
f014b69495
|
Add tests for the sample_interval in #3069
|
2014-04-17 00:57:26 +02:00 |
|
Lars Buitinck
|
61142601d8
|
FIX input validation in Nystroem
Minimal, so we can later add support for dtype=object and custom kernels
on strings, trees, graphs, etc.
|
2014-03-05 13:14:46 +01:00 |
|
Lars Buitinck
|
9a9b9e85ed
|
DOC+COSMIT: typos, lots of them
Fixes #2102. There's one typo left in joblib which is better fixed
upstream.
|
2013-09-13 16:29:42 +02:00 |
|
Lars Buitinck
|
b41eae2318
|
COSMIT rm repetitive __main__ blocks from tests
These may seem convenient, but they have to updated along with the tests
and can give a false sense of security when that doesn't happen.
|
2013-03-18 15:28:46 +01:00 |
|
JakeMick
|
5f6c0db2db
|
TST added test of fit and transform for kernels for nystroem
|
2013-03-18 12:51:57 +01:00 |
|
Lars Buitinck
|
66eb410bdd
|
BUG + DOC fix Nystroem for other kernels than RBF
Fixes #1785, #1786.
|
2013-03-18 12:51:50 +01:00 |
|
Andreas Mueller
|
f52e8116b8
|
TST better coverage in kernel_approximation
|
2012-12-12 22:09:08 +01:00 |
|
Andreas Mueller
|
f3e8fb840e
|
ENH tests for Nystrom, check that n_components is smaller than n_samples.
|
2012-12-09 14:14:09 +01:00 |
|
Mathieu Blondel
|
afb666c1d0
|
Use sklearn.utils.testing in tests.
|
2012-10-25 15:31:35 +09:00 |
|
Lars Buitinck
|
11e9842987
|
ENH scipy.sparse support in additive chi² sampler
|
2012-06-23 14:04:04 +02:00 |
|
Andreas Mueller
|
eeb6dcc4cc
|
ENH use random states everywhere, never call np.random.
|
2012-05-06 18:35:14 +02:00 |
|
Mathieu Blondel
|
e3ed8e0a0e
|
Use rbf_kernel.
|
2012-01-10 13:35:15 +09:00 |
|
Lars Buitinck
|
872e70d9b7
|
BUG input validation in kernel approximations + pep8
Note: SkewedChi2Sampler no longer accepts sparse matrices.
It shouldn't, because it destroyed the sparseness anyway.
|
2011-12-24 21:21:23 +01:00 |
|
Andreas Mueller
|
3a6e490e05
|
COSMIT pep8
|
2011-12-23 19:12:52 +01:00 |
|
Andreas Mueller
|
48be642417
|
ENH better testing
|
2011-12-20 14:17:50 +01:00 |
|
Andreas Mueller
|
7525923d13
|
ENH Addressing some of Gael's comments, mainly naming and docstrings
|
2011-12-20 14:17:50 +01:00 |
|
Andreas Mueller
|
972de6d452
|
Moving kernel_approximation to the top level
|
2011-12-20 14:17:49 +01:00 |