679936f6bc
MAINT Deprecate LDA/QDA in favor of expanded names
...
The LDA accronym for Linear Discriminant Analysis is ambiguous
because of the newly introduced Latent Dirichlet Allocation model.
We therefore deprecate the sklearn.lda.LDA and sklearn.lda.QDA
in favor of explicit names.
2015-09-11 15:32:11 +02:00
9ebc6fe33d
Merge pull request #5216 from JPFrancoia/master
...
Creation of the attribute LDA.explained_variance_ratio_, for the eige…
2015-09-11 11:42:39 +02:00
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
6e7e9d3df5
Modification of a line after @agramfort's suggestion.
2015-09-06 17:21:47 +02:00
cf4bc71caa
Changing [ to (. Formatting consistency.
2015-09-06 17:19:05 +02:00
3ad9972192
Addition of a piece of docstring for the attribute
...
explained_variance_ratio_. Creation of the function
test_lda_explained_variance_ratio to test this new attribute.
2015-09-06 11:46:54 +02:00
5a1f18312b
Creation of the attribute LDA.explained_variance_ratio_, for the eigen solver.
...
It is an analogy to PCA.explained_variance_ratio, and works in the same
way.
2015-09-05 17:44:05 +02:00
7befb5d970
Explicit broadcasting
2015-08-06 17:34:46 +02:00
1d98ca368a
Fixed shrinkage covariance for LDA
2015-08-05 16:32:38 +02:00
0650d5502e
DOC adding backlinks to docstrings
2015-06-03 00:24:04 -04:00
c02cc3dd23
FIX LDA(solver="lsqr"): make sure the right error is raised on transform.
2015-03-20 12:27:24 -04:00
a13a5859e8
DOC remove sphinx warnings when generating the doc
...
and fixes a few other doc problems while I was at it.
The main take-home message is to remember that you need a backslash when
your parameter type description exceeds one line.
2015-03-18 15:49:30 +01:00
67dcd99d82
TST fix tests with numpy 1.6.1
...
np.bincount raises an Exception with empty input arrays for numpy versions < 1.6.2
Add utils.fixes.bincount to tackle this issue and use it instead of np.bincount
2015-02-01 15:49:49 +01:00
cb77255ee6
fix import error in lda.py
2015-01-18 00:32:19 -05:00
a413f875f1
DOC minor fixes do docstrings, don't document deprecated parameters.
2015-01-16 11:55:02 -05:00
b3fbccca38
FIX various mismatch between docstring and signature params
...
DOC max_iterations -> max_iter. Make it consistent with kmeans
MAINT Replace the deprecated dx parameter with d in the docstrings
MAINT Deprecation warning for max_iterations parameter.
2015-01-16 11:54:57 -05:00
abd31d2cad
MAINT Make uniform the error raised for not fitted condition
2015-01-12 00:33:40 +05:30
46c1c42131
remove deprecated ``n_bootstraps`` parameter in Bootstrap cv object, verbose in FactorAnalysis and SelectorMixin. Also adjust some deprecation strings.
2014-12-29 13:53:42 -05:00
081418f573
Added documentation for the three solvers.
2014-12-16 11:50:02 +01:00
71a23efee9
Fix problem with np.linalg.norm for version <= 1.6.
2014-12-16 11:50:02 +01:00
936acf7e1f
Works with one feature.
2014-12-16 11:50:02 +01:00
095c7a90c6
Now correctly use weighted average to compute the class covariance. Cleaned up code and added references.
2014-12-16 11:50:02 +01:00
70773bc5bc
Added deprecation warnings to store_covariance and tol parameters in fit() method.
2014-12-16 11:50:01 +01:00
aede3d64f1
Rename parameter value 'ledoit_wolf' to 'auto'.
2014-12-16 11:50:01 +01:00
7e4356b940
Inherit from LinearClassifierMixin.
2014-12-16 11:35:59 +01:00
f2fcafebcb
Another try to fix the Travis bug.
2014-12-16 11:35:59 +01:00
7a5563d128
(Hopefully) fixed Travis bug on Python 2.6.
2014-12-16 11:35:59 +01:00
7c1101d7c2
Treat the binary case differently from the case when n_classes > 2.
2014-12-16 11:35:59 +01:00
9d9eb91062
Changed parametrization of coef_ and intercept_.
2014-12-16 11:35:59 +01:00
a5ae150846
rename alpha to shrinkage + docstring cosmits
2014-12-16 11:35:58 +01:00
5deb80f7b1
Fixed docstrings.
2014-12-16 11:35:58 +01:00
97ab90b5b2
By default, estimate priors from sample.
2014-12-16 11:35:58 +01:00
685c62794e
Use broadcasting to scale back instead of multiplying with a diagonal matrix.
2014-12-16 11:35:58 +01:00
764efbe605
Changed format of shape parameter in docstrings.
2014-12-16 11:35:58 +01:00
76963c0712
Addressed some comments by @agramfort.
2014-12-16 11:35:58 +01:00
4ba7c3aa43
Cleaned up one more docstring.
2014-12-16 11:35:57 +01:00
84a41f5533
Cleaned up docstrings.
2014-12-16 11:35:57 +01:00
4dede9aedf
Raise error if alpha is not the expected type
2014-12-16 11:35:57 +01:00
1ff5199248
Replaced str with six.string_types
2014-12-16 11:35:57 +01:00
b787f8038f
Revert last commit because it broke some tests
2014-12-16 11:35:57 +01:00
4140249893
Removed private _decision_function()
2014-12-16 11:35:57 +01:00
e4f4f6df1b
Fixed doctest
2014-12-16 11:35:57 +01:00
c93485f059
PEP8 style conventions
2014-12-16 11:35:56 +01:00
6296194edd
Parameters are now consistently named; also refactored the _means_cov() function
2014-12-16 11:35:56 +01:00
b9601ccf94
Addressed @mblondel's comments
2014-12-16 11:35:56 +01:00
7216260944
For now only the svd solver supports the transform method
2014-12-16 11:35:55 +01:00
27f696d9fa
Support manually set shrinkage parameter (alpha)
2014-12-16 11:35:55 +01:00
a98e0d361d
Moved some common code into helper function
2014-12-16 11:35:55 +01:00
f130922b2e
Use equal priors by default
2014-12-16 11:35:55 +01:00
981fbe0608
Return self in fit()
2014-12-16 11:35:55 +01:00
Vinayak Mehta