5db2adf93c
MAINT Removed deprecated stuff.
2015-10-21 10:24:23 +02:00
a87011262a
FIX precision to float64 across the codebase
2015-10-16 11:20:29 +02:00
c805fbc79b
Refactor tests
2015-10-10 22:05:40 -04:00
5a0db1717a
1. Added parameter prefit to pass in a fitted estimator.
...
2. Use assert_warns instead of catch_warnings
3. Remove depracation warnings in common tests.
2015-10-09 01:57:36 -04:00
acf5f160a8
Merge SelectFromModel and L1-selection examples
...
Add test to check the threshold can be set without refitting
2015-10-05 11:25:07 -04:00
2ee718cc75
Add narrative docs and fix examples
2015-10-05 11:25:06 -04:00
10176d9897
Now a fitted estimator can be passed to SelectFromModel
2015-10-05 11:25:06 -04:00
459cb9ba3d
Remove warm start
2015-10-05 11:19:57 -04:00
c438f78996
Implemented SelectFromModel meta-transformer
2015-10-05 11:19:57 -04:00
2f099334f9
MAINT deprecate 1d input arrays for all estimators
...
Passing 1D arrays to check_array, without setting `ensure_2d` to false now
raises a deprecation warning before reshaping it. This will later throw an
error.
All Scaler classes also throw warnings when 1D arrays are passed.
All unit tests/doctests are modified to ensure that no 1D arrays are passed,
except in explicit 1D array tests where the warnings have been silenced.
Additional tests are also included which check for different 1D array cases.
2D array tests with one samples and one features are also added and where
they failed, `check_array` call has been modified to give a more useful error
message
2015-09-09 15:49:58 +02:00
f369d3ab0f
Fix RFE / RFECV estimator tags
2015-05-29 11:36:38 -04:00
79ea587e05
FIX make rfe feature importance test deterministic.
2015-04-28 18:26:22 -04:00
c165327940
Add regression test for the number of subsets of features
2015-04-20 12:58:29 -04:00
7bbf10a598
ENH optimize rfecv by eliminating the for loop
2015-04-20 12:27:31 -04:00
0f0e58d35e
Fix RFE - Add the possibility to use feature_importances rather than coef_ when existing, see #2121
2015-04-02 14:56:48 +02:00
17ee7dd79b
DOC & COSMIT deprecate estimator_params in RFE and RFE_CV in docstring
2015-03-29 12:47:53 -04:00
cd2ee7e454
MAINT docstring --> comments to prevent nose from using doc in verbose mode
2015-03-21 11:16:49 +05:30
b741aa3e5c
FIX coding style #4292
2015-03-10 14:46:10 -04:00
9f312a72b3
Correct warning messages and move warnings to fit method #4292
2015-03-10 10:47:15 -04:00
08b97b32c2
Deprecate estimator_params in RFE and RFECV #4292
2015-03-09 16:45:46 -04:00
f12a441655
raise noise in test again, be ascii
2015-02-24 16:26:24 -05:00
95f1d9562b
FIX DOC TST: alpha is an uppper bound on FDR
2015-02-24 16:26:24 -05:00
700babc43f
extensive (excessive?) testing of FDR
2015-02-24 16:26:24 -05:00
1add39ff11
[Feature Selection] Fix SelectFDR thresholding bug ( #2771 )
...
From https://github.com/scikit-learn/scikit-learn/issues/2771 , we were
not correctly scaling the alpha
parameter (http://en.wikipedia.org/wiki/False_discovery_rate#Benjamini.E2.80.93Hochberg_procedure )
with the number of features (== hypothesis). Thus, the alpha
parameter was not invariant wrt the number of features.
The correction is as suggested in the original issue, and a test has
been added that verifies that for various numbers of features, an
appropriate false discovery rate is generated when using the selector.
2015-02-24 16:25:58 -05:00
5c0c5c672c
MAINT use absolute imports in tests
...
as per the guideline in:
http://scikit-learn.org/stable/developers/#coding-guidelines
2015-02-10 13:40:31 +01:00
95681eecca
Merge pull request #4206 from ogrisel/fix-strict-select-fdr
...
[MRG] explicit warning message for strict selectors
Also fixes #4059
2015-02-07 21:54:45 +11:00
2ce9eac717
TST more strict selectors with empty support
2015-02-07 11:37:06 +01:00
bc5b403af3
FIX #4059 : explicit warning for strict selectors
2015-02-06 12:06:34 +01:00
97e3b96801
PEP8 compliance
2015-02-05 15:26:46 -05:00
13b6aec5f3
non-inheriting estimator added for rfe, cross validation, and pipeline tests
...
non-inheriting estimator added for cross validation
non-inheriting estimator added for pipeline
reconciled diffs
2015-01-29 22:02:38 -05:00
88555987cc
PEP8 Fix E101, E111 errors and W191, W293, W293 and W391 warnings.
2015-01-08 04:28:24 +05:30
6039b1864b
Merge pull request #3824 from nmayorov/rfecv_bugfix
...
[MRG+1] FIX: Bug in RFECV when step is not 1
2014-12-15 09:06:41 +01:00
ad7e772422
Fixed a bug in RFECV when step != 1
2014-12-15 02:53:27 +03:00
081a554698
FIX P/R/F metrics and scorers are now for binary problems only by default
...
Scorers for different average parameters have been added.
2014-12-09 13:56:51 +11:00
2fecf6facf
Fix RFE n_features minimum value #3812 and make it simpler.
2014-11-24 18:05:00 +01:00
60cadce9d1
Fix RFE n_features minimum value #3812
2014-11-20 19:19:40 +01:00
0bd095f9d6
FIX: Raise warnings in f_classif a given feature is constant throughout
2014-10-08 19:32:47 +02:00
e01749068c
MAINT fix prng in test_f_oneway_ints
...
Not fixing the PRNG cause the test to fail randomly under windows
where the 4 decimals place check can be too strict for some unlucky
random draws of test data.
2014-07-29 08:54:47 +02:00
f3afd4e5e2
FIX unstable test on 32 bit windows
2014-07-23 19:11:53 +02:00
6e2a83b4e1
remove check_arrays stuff and old input validation
2014-07-20 13:31:45 +02:00
b12d1ac487
Remove deprecated 'score_func' and 'loss_func' parameters from sklearn.metrics.scorer.check_scoring. Amend the code in all the other places they were used.
2014-07-17 13:53:06 +01:00
9248948344
test case on rtfe grid_scores fix
2014-07-15 16:35:04 +02:00
724b63a16c
Replacing the test .todense() methods with .toarray()
2014-05-20 21:44:34 -04:00
db4230a776
TST improve parameter checking + use nose assert_ functions
2014-05-08 20:56:21 +02:00
e58405d6f6
MAINT remove redundant class hierarchy + fix api perform parameter check in fit
2014-05-08 20:56:21 +02:00
a3f81d9a04
Merge pull request #3010 from griffinmyers/fix-select-k-best
...
Fixed SelectKBest corner case: k=0
2014-03-28 09:32:31 +11:00
f1e2e1c874
Fixed SelectKBest corner case: k=0
...
Without this fix k = 0 will have the same behavior as k = 'all' based on how
the array indexing was written.
I've included a test that demonstrates proper behaviour.
2014-03-27 15:18:50 -04:00
2b2c5b2db8
FIX: np.random.randint expects signed 32 bit integers under Windows
2014-03-27 14:20:03 +01:00
4c129e57c1
f_regression and degress of freedom: update whats_new + minor.
2014-02-07 19:28:46 +01:00
4d18fd8837
BF in f_regression: variable naming + use assert_*array*_almost_equal.
2014-02-07 19:28:06 +01:00
Arnaud Rachez