forked from ukosuagwu/scikit-learn
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy path_split.py
More file actions
2714 lines (2249 loc) · 93.4 KB
/
Copy path_split.py
File metadata and controls
2714 lines (2249 loc) · 93.4 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
"""
The :mod:`sklearn.model_selection._split` module includes classes and
functions to split the data based on a preset strategy.
"""
# Author: Alexandre Gramfort <alexandre.gramfort@inria.fr>
# Gael Varoquaux <gael.varoquaux@normalesup.org>
# Olivier Grisel <olivier.grisel@ensta.org>
# Raghav RV <rvraghav93@gmail.com>
# Leandro Hermida <hermidal@cs.umd.edu>
# Rodion Martynov <marrodion@gmail.com>
# License: BSD 3 clause
from collections.abc import Iterable
from collections import defaultdict
import warnings
from itertools import chain, combinations
from math import ceil, floor
import numbers
from abc import ABCMeta, abstractmethod
from inspect import signature
import numpy as np
from scipy.special import comb
from ..utils import indexable, check_random_state, _safe_indexing
from ..utils import _approximate_mode
from ..utils.validation import _num_samples, column_or_1d
from ..utils.validation import check_array
from ..utils.multiclass import type_of_target
from ..utils._param_validation import validate_params, Interval
__all__ = [
"BaseCrossValidator",
"KFold",
"GroupKFold",
"LeaveOneGroupOut",
"LeaveOneOut",
"LeavePGroupsOut",
"LeavePOut",
"RepeatedStratifiedKFold",
"RepeatedKFold",
"ShuffleSplit",
"GroupShuffleSplit",
"StratifiedKFold",
"StratifiedGroupKFold",
"StratifiedShuffleSplit",
"PredefinedSplit",
"train_test_split",
"check_cv",
]
class BaseCrossValidator(metaclass=ABCMeta):
"""Base class for all cross-validators
Implementations must define `_iter_test_masks` or `_iter_test_indices`.
"""
def split(self, X, y=None, groups=None):
"""Generate indices to split data into training and test set.
Parameters
----------
X : array-like of shape (n_samples, n_features)
Training data, where `n_samples` is the number of samples
and `n_features` is the number of features.
y : array-like of shape (n_samples,)
The target variable for supervised learning problems.
groups : array-like of shape (n_samples,), default=None
Group labels for the samples used while splitting the dataset into
train/test set.
Yields
------
train : ndarray
The training set indices for that split.
test : ndarray
The testing set indices for that split.
"""
X, y, groups = indexable(X, y, groups)
indices = np.arange(_num_samples(X))
for test_index in self._iter_test_masks(X, y, groups):
train_index = indices[np.logical_not(test_index)]
test_index = indices[test_index]
yield train_index, test_index
# Since subclasses must implement either _iter_test_masks or
# _iter_test_indices, neither can be abstract.
def _iter_test_masks(self, X=None, y=None, groups=None):
"""Generates boolean masks corresponding to test sets.
By default, delegates to _iter_test_indices(X, y, groups)
"""
for test_index in self._iter_test_indices(X, y, groups):
test_mask = np.zeros(_num_samples(X), dtype=bool)
test_mask[test_index] = True
yield test_mask
def _iter_test_indices(self, X=None, y=None, groups=None):
"""Generates integer indices corresponding to test sets."""
raise NotImplementedError
@abstractmethod
def get_n_splits(self, X=None, y=None, groups=None):
"""Returns the number of splitting iterations in the cross-validator"""
def __repr__(self):
return _build_repr(self)
class LeaveOneOut(BaseCrossValidator):
"""Leave-One-Out cross-validator
Provides train/test indices to split data in train/test sets. Each
sample is used once as a test set (singleton) while the remaining
samples form the training set.
Note: ``LeaveOneOut()`` is equivalent to ``KFold(n_splits=n)`` and
``LeavePOut(p=1)`` where ``n`` is the number of samples.
Due to the high number of test sets (which is the same as the
number of samples) this cross-validation method can be very costly.
For large datasets one should favor :class:`KFold`, :class:`ShuffleSplit`
or :class:`StratifiedKFold`.
Read more in the :ref:`User Guide <leave_one_out>`.
Examples
--------
>>> import numpy as np
>>> from sklearn.model_selection import LeaveOneOut
>>> X = np.array([[1, 2], [3, 4]])
>>> y = np.array([1, 2])
>>> loo = LeaveOneOut()
>>> loo.get_n_splits(X)
2
>>> print(loo)
LeaveOneOut()
>>> for i, (train_index, test_index) in enumerate(loo.split(X)):
... print(f"Fold {i}:")
... print(f" Train: index={train_index}")
... print(f" Test: index={test_index}")
Fold 0:
Train: index=[1]
Test: index=[0]
Fold 1:
Train: index=[0]
Test: index=[1]
See Also
--------
LeaveOneGroupOut : For splitting the data according to explicit,
domain-specific stratification of the dataset.
GroupKFold : K-fold iterator variant with non-overlapping groups.
"""
def _iter_test_indices(self, X, y=None, groups=None):
n_samples = _num_samples(X)
if n_samples <= 1:
raise ValueError(
"Cannot perform LeaveOneOut with n_samples={}.".format(n_samples)
)
return range(n_samples)
def get_n_splits(self, X, y=None, groups=None):
"""Returns the number of splitting iterations in the cross-validator
Parameters
----------
X : array-like of shape (n_samples, n_features)
Training data, where `n_samples` is the number of samples
and `n_features` is the number of features.
y : object
Always ignored, exists for compatibility.
groups : object
Always ignored, exists for compatibility.
Returns
-------
n_splits : int
Returns the number of splitting iterations in the cross-validator.
"""
if X is None:
raise ValueError("The 'X' parameter should not be None.")
return _num_samples(X)
class LeavePOut(BaseCrossValidator):
"""Leave-P-Out cross-validator
Provides train/test indices to split data in train/test sets. This results
in testing on all distinct samples of size p, while the remaining n - p
samples form the training set in each iteration.
Note: ``LeavePOut(p)`` is NOT equivalent to
``KFold(n_splits=n_samples // p)`` which creates non-overlapping test sets.
Due to the high number of iterations which grows combinatorically with the
number of samples this cross-validation method can be very costly. For
large datasets one should favor :class:`KFold`, :class:`StratifiedKFold`
or :class:`ShuffleSplit`.
Read more in the :ref:`User Guide <leave_p_out>`.
Parameters
----------
p : int
Size of the test sets. Must be strictly less than the number of
samples.
Examples
--------
>>> import numpy as np
>>> from sklearn.model_selection import LeavePOut
>>> X = np.array([[1, 2], [3, 4], [5, 6], [7, 8]])
>>> y = np.array([1, 2, 3, 4])
>>> lpo = LeavePOut(2)
>>> lpo.get_n_splits(X)
6
>>> print(lpo)
LeavePOut(p=2)
>>> for i, (train_index, test_index) in enumerate(lpo.split(X)):
... print(f"Fold {i}:")
... print(f" Train: index={train_index}")
... print(f" Test: index={test_index}")
Fold 0:
Train: index=[2 3]
Test: index=[0 1]
Fold 1:
Train: index=[1 3]
Test: index=[0 2]
Fold 2:
Train: index=[1 2]
Test: index=[0 3]
Fold 3:
Train: index=[0 3]
Test: index=[1 2]
Fold 4:
Train: index=[0 2]
Test: index=[1 3]
Fold 5:
Train: index=[0 1]
Test: index=[2 3]
"""
def __init__(self, p):
self.p = p
def _iter_test_indices(self, X, y=None, groups=None):
n_samples = _num_samples(X)
if n_samples <= self.p:
raise ValueError(
"p={} must be strictly less than the number of samples={}".format(
self.p, n_samples
)
)
for combination in combinations(range(n_samples), self.p):
yield np.array(combination)
def get_n_splits(self, X, y=None, groups=None):
"""Returns the number of splitting iterations in the cross-validator
Parameters
----------
X : array-like of shape (n_samples, n_features)
Training data, where `n_samples` is the number of samples
and `n_features` is the number of features.
y : object
Always ignored, exists for compatibility.
groups : object
Always ignored, exists for compatibility.
"""
if X is None:
raise ValueError("The 'X' parameter should not be None.")
return int(comb(_num_samples(X), self.p, exact=True))
class _BaseKFold(BaseCrossValidator, metaclass=ABCMeta):
"""Base class for KFold, GroupKFold, and StratifiedKFold"""
@abstractmethod
def __init__(self, n_splits, *, shuffle, random_state):
if not isinstance(n_splits, numbers.Integral):
raise ValueError(
"The number of folds must be of Integral type. "
"%s of type %s was passed." % (n_splits, type(n_splits))
)
n_splits = int(n_splits)
if n_splits <= 1:
raise ValueError(
"k-fold cross-validation requires at least one"
" train/test split by setting n_splits=2 or more,"
" got n_splits={0}.".format(n_splits)
)
if not isinstance(shuffle, bool):
raise TypeError("shuffle must be True or False; got {0}".format(shuffle))
if not shuffle and random_state is not None: # None is the default
raise ValueError(
"Setting a random_state has no effect since shuffle is "
"False. You should leave "
"random_state to its default (None), or set shuffle=True.",
)
self.n_splits = n_splits
self.shuffle = shuffle
self.random_state = random_state
def split(self, X, y=None, groups=None):
"""Generate indices to split data into training and test set.
Parameters
----------
X : array-like of shape (n_samples, n_features)
Training data, where `n_samples` is the number of samples
and `n_features` is the number of features.
y : array-like of shape (n_samples,), default=None
The target variable for supervised learning problems.
groups : array-like of shape (n_samples,), default=None
Group labels for the samples used while splitting the dataset into
train/test set.
Yields
------
train : ndarray
The training set indices for that split.
test : ndarray
The testing set indices for that split.
"""
X, y, groups = indexable(X, y, groups)
n_samples = _num_samples(X)
if self.n_splits > n_samples:
raise ValueError(
(
"Cannot have number of splits n_splits={0} greater"
" than the number of samples: n_samples={1}."
).format(self.n_splits, n_samples)
)
for train, test in super().split(X, y, groups):
yield train, test
def get_n_splits(self, X=None, y=None, groups=None):
"""Returns the number of splitting iterations in the cross-validator
Parameters
----------
X : object
Always ignored, exists for compatibility.
y : object
Always ignored, exists for compatibility.
groups : object
Always ignored, exists for compatibility.
Returns
-------
n_splits : int
Returns the number of splitting iterations in the cross-validator.
"""
return self.n_splits
class KFold(_BaseKFold):
"""K-Folds cross-validator
Provides train/test indices to split data in train/test sets. Split
dataset into k consecutive folds (without shuffling by default).
Each fold is then used once as a validation while the k - 1 remaining
folds form the training set.
Read more in the :ref:`User Guide <k_fold>`.
Parameters
----------
n_splits : int, default=5
Number of folds. Must be at least 2.
.. versionchanged:: 0.22
``n_splits`` default value changed from 3 to 5.
shuffle : bool, default=False
Whether to shuffle the data before splitting into batches.
Note that the samples within each split will not be shuffled.
random_state : int, RandomState instance or None, default=None
When `shuffle` is True, `random_state` affects the ordering of the
indices, which controls the randomness of each fold. Otherwise, this
parameter has no effect.
Pass an int for reproducible output across multiple function calls.
See :term:`Glossary <random_state>`.
Examples
--------
>>> import numpy as np
>>> from sklearn.model_selection import KFold
>>> X = np.array([[1, 2], [3, 4], [1, 2], [3, 4]])
>>> y = np.array([1, 2, 3, 4])
>>> kf = KFold(n_splits=2)
>>> kf.get_n_splits(X)
2
>>> print(kf)
KFold(n_splits=2, random_state=None, shuffle=False)
>>> for i, (train_index, test_index) in enumerate(kf.split(X)):
... print(f"Fold {i}:")
... print(f" Train: index={train_index}")
... print(f" Test: index={test_index}")
Fold 0:
Train: index=[2 3]
Test: index=[0 1]
Fold 1:
Train: index=[0 1]
Test: index=[2 3]
Notes
-----
The first ``n_samples % n_splits`` folds have size
``n_samples // n_splits + 1``, other folds have size
``n_samples // n_splits``, where ``n_samples`` is the number of samples.
Randomized CV splitters may return different results for each call of
split. You can make the results identical by setting `random_state`
to an integer.
See Also
--------
StratifiedKFold : Takes class information into account to avoid building
folds with imbalanced class distributions (for binary or multiclass
classification tasks).
GroupKFold : K-fold iterator variant with non-overlapping groups.
RepeatedKFold : Repeats K-Fold n times.
"""
def __init__(self, n_splits=5, *, shuffle=False, random_state=None):
super().__init__(n_splits=n_splits, shuffle=shuffle, random_state=random_state)
def _iter_test_indices(self, X, y=None, groups=None):
n_samples = _num_samples(X)
indices = np.arange(n_samples)
if self.shuffle:
check_random_state(self.random_state).shuffle(indices)
n_splits = self.n_splits
fold_sizes = np.full(n_splits, n_samples // n_splits, dtype=int)
fold_sizes[: n_samples % n_splits] += 1
current = 0
for fold_size in fold_sizes:
start, stop = current, current + fold_size
yield indices[start:stop]
current = stop
class GroupKFold(_BaseKFold):
"""K-fold iterator variant with non-overlapping groups.
Each group will appear exactly once in the test set across all folds (the
number of distinct groups has to be at least equal to the number of folds).
The folds are approximately balanced in the sense that the number of
distinct groups is approximately the same in each fold.
Read more in the :ref:`User Guide <group_k_fold>`.
Parameters
----------
n_splits : int, default=5
Number of folds. Must be at least 2.
.. versionchanged:: 0.22
``n_splits`` default value changed from 3 to 5.
Notes
-----
Groups appear in an arbitrary order throughout the folds.
Examples
--------
>>> import numpy as np
>>> from sklearn.model_selection import GroupKFold
>>> X = np.array([[1, 2], [3, 4], [5, 6], [7, 8], [9, 10], [11, 12]])
>>> y = np.array([1, 2, 3, 4, 5, 6])
>>> groups = np.array([0, 0, 2, 2, 3, 3])
>>> group_kfold = GroupKFold(n_splits=2)
>>> group_kfold.get_n_splits(X, y, groups)
2
>>> print(group_kfold)
GroupKFold(n_splits=2)
>>> for i, (train_index, test_index) in enumerate(group_kfold.split(X, y, groups)):
... print(f"Fold {i}:")
... print(f" Train: index={train_index}, group={groups[train_index]}")
... print(f" Test: index={test_index}, group={groups[test_index]}")
Fold 0:
Train: index=[2 3], group=[2 2]
Test: index=[0 1 4 5], group=[0 0 3 3]
Fold 1:
Train: index=[0 1 4 5], group=[0 0 3 3]
Test: index=[2 3], group=[2 2]
See Also
--------
LeaveOneGroupOut : For splitting the data according to explicit
domain-specific stratification of the dataset.
StratifiedKFold : Takes class information into account to avoid building
folds with imbalanced class proportions (for binary or multiclass
classification tasks).
"""
def __init__(self, n_splits=5):
super().__init__(n_splits, shuffle=False, random_state=None)
def _iter_test_indices(self, X, y, groups):
if groups is None:
raise ValueError("The 'groups' parameter should not be None.")
groups = check_array(groups, input_name="groups", ensure_2d=False, dtype=None)
unique_groups, groups = np.unique(groups, return_inverse=True)
n_groups = len(unique_groups)
if self.n_splits > n_groups:
raise ValueError(
"Cannot have number of splits n_splits=%d greater"
" than the number of groups: %d." % (self.n_splits, n_groups)
)
# Weight groups by their number of occurrences
n_samples_per_group = np.bincount(groups)
# Distribute the most frequent groups first
indices = np.argsort(n_samples_per_group)[::-1]
n_samples_per_group = n_samples_per_group[indices]
# Total weight of each fold
n_samples_per_fold = np.zeros(self.n_splits)
# Mapping from group index to fold index
group_to_fold = np.zeros(len(unique_groups))
# Distribute samples by adding the largest weight to the lightest fold
for group_index, weight in enumerate(n_samples_per_group):
lightest_fold = np.argmin(n_samples_per_fold)
n_samples_per_fold[lightest_fold] += weight
group_to_fold[indices[group_index]] = lightest_fold
indices = group_to_fold[groups]
for f in range(self.n_splits):
yield np.where(indices == f)[0]
def split(self, X, y=None, groups=None):
"""Generate indices to split data into training and test set.
Parameters
----------
X : array-like of shape (n_samples, n_features)
Training data, where `n_samples` is the number of samples
and `n_features` is the number of features.
y : array-like of shape (n_samples,), default=None
The target variable for supervised learning problems.
groups : array-like of shape (n_samples,)
Group labels for the samples used while splitting the dataset into
train/test set.
Yields
------
train : ndarray
The training set indices for that split.
test : ndarray
The testing set indices for that split.
"""
return super().split(X, y, groups)
class StratifiedKFold(_BaseKFold):
"""Stratified K-Folds cross-validator.
Provides train/test indices to split data in train/test sets.
This cross-validation object is a variation of KFold that returns
stratified folds. The folds are made by preserving the percentage of
samples for each class.
Read more in the :ref:`User Guide <stratified_k_fold>`.
Parameters
----------
n_splits : int, default=5
Number of folds. Must be at least 2.
.. versionchanged:: 0.22
``n_splits`` default value changed from 3 to 5.
shuffle : bool, default=False
Whether to shuffle each class's samples before splitting into batches.
Note that the samples within each split will not be shuffled.
random_state : int, RandomState instance or None, default=None
When `shuffle` is True, `random_state` affects the ordering of the
indices, which controls the randomness of each fold for each class.
Otherwise, leave `random_state` as `None`.
Pass an int for reproducible output across multiple function calls.
See :term:`Glossary <random_state>`.
Examples
--------
>>> import numpy as np
>>> from sklearn.model_selection import StratifiedKFold
>>> X = np.array([[1, 2], [3, 4], [1, 2], [3, 4]])
>>> y = np.array([0, 0, 1, 1])
>>> skf = StratifiedKFold(n_splits=2)
>>> skf.get_n_splits(X, y)
2
>>> print(skf)
StratifiedKFold(n_splits=2, random_state=None, shuffle=False)
>>> for i, (train_index, test_index) in enumerate(skf.split(X, y)):
... print(f"Fold {i}:")
... print(f" Train: index={train_index}")
... print(f" Test: index={test_index}")
Fold 0:
Train: index=[1 3]
Test: index=[0 2]
Fold 1:
Train: index=[0 2]
Test: index=[1 3]
Notes
-----
The implementation is designed to:
* Generate test sets such that all contain the same distribution of
classes, or as close as possible.
* Be invariant to class label: relabelling ``y = ["Happy", "Sad"]`` to
``y = [1, 0]`` should not change the indices generated.
* Preserve order dependencies in the dataset ordering, when
``shuffle=False``: all samples from class k in some test set were
contiguous in y, or separated in y by samples from classes other than k.
* Generate test sets where the smallest and largest differ by at most one
sample.
.. versionchanged:: 0.22
The previous implementation did not follow the last constraint.
See Also
--------
RepeatedStratifiedKFold : Repeats Stratified K-Fold n times.
"""
def __init__(self, n_splits=5, *, shuffle=False, random_state=None):
super().__init__(n_splits=n_splits, shuffle=shuffle, random_state=random_state)
def _make_test_folds(self, X, y=None):
rng = check_random_state(self.random_state)
y = np.asarray(y)
type_of_target_y = type_of_target(y)
allowed_target_types = ("binary", "multiclass")
if type_of_target_y not in allowed_target_types:
raise ValueError(
"Supported target types are: {}. Got {!r} instead.".format(
allowed_target_types, type_of_target_y
)
)
y = column_or_1d(y)
_, y_idx, y_inv = np.unique(y, return_index=True, return_inverse=True)
# y_inv encodes y according to lexicographic order. We invert y_idx to
# map the classes so that they are encoded by order of appearance:
# 0 represents the first label appearing in y, 1 the second, etc.
_, class_perm = np.unique(y_idx, return_inverse=True)
y_encoded = class_perm[y_inv]
n_classes = len(y_idx)
y_counts = np.bincount(y_encoded)
min_groups = np.min(y_counts)
if np.all(self.n_splits > y_counts):
raise ValueError(
"n_splits=%d cannot be greater than the"
" number of members in each class." % (self.n_splits)
)
if self.n_splits > min_groups:
warnings.warn(
"The least populated class in y has only %d"
" members, which is less than n_splits=%d."
% (min_groups, self.n_splits),
UserWarning,
)
# Determine the optimal number of samples from each class in each fold,
# using round robin over the sorted y. (This can be done direct from
# counts, but that code is unreadable.)
y_order = np.sort(y_encoded)
allocation = np.asarray(
[
np.bincount(y_order[i :: self.n_splits], minlength=n_classes)
for i in range(self.n_splits)
]
)
# To maintain the data order dependencies as best as possible within
# the stratification constraint, we assign samples from each class in
# blocks (and then mess that up when shuffle=True).
test_folds = np.empty(len(y), dtype="i")
for k in range(n_classes):
# since the kth column of allocation stores the number of samples
# of class k in each test set, this generates blocks of fold
# indices corresponding to the allocation for class k.
folds_for_class = np.arange(self.n_splits).repeat(allocation[:, k])
if self.shuffle:
rng.shuffle(folds_for_class)
test_folds[y_encoded == k] = folds_for_class
return test_folds
def _iter_test_masks(self, X, y=None, groups=None):
test_folds = self._make_test_folds(X, y)
for i in range(self.n_splits):
yield test_folds == i
def split(self, X, y, groups=None):
"""Generate indices to split data into training and test set.
Parameters
----------
X : array-like of shape (n_samples, n_features)
Training data, where `n_samples` is the number of samples
and `n_features` is the number of features.
Note that providing ``y`` is sufficient to generate the splits and
hence ``np.zeros(n_samples)`` may be used as a placeholder for
``X`` instead of actual training data.
y : array-like of shape (n_samples,)
The target variable for supervised learning problems.
Stratification is done based on the y labels.
groups : object
Always ignored, exists for compatibility.
Yields
------
train : ndarray
The training set indices for that split.
test : ndarray
The testing set indices for that split.
Notes
-----
Randomized CV splitters may return different results for each call of
split. You can make the results identical by setting `random_state`
to an integer.
"""
y = check_array(y, input_name="y", ensure_2d=False, dtype=None)
return super().split(X, y, groups)
class StratifiedGroupKFold(_BaseKFold):
"""Stratified K-Folds iterator variant with non-overlapping groups.
This cross-validation object is a variation of StratifiedKFold attempts to
return stratified folds with non-overlapping groups. The folds are made by
preserving the percentage of samples for each class.
Each group will appear exactly once in the test set across all folds (the
number of distinct groups has to be at least equal to the number of folds).
The difference between :class:`~sklearn.model_selection.GroupKFold`
and :class:`~sklearn.model_selection.StratifiedGroupKFold` is that
the former attempts to create balanced folds such that the number of
distinct groups is approximately the same in each fold, whereas
StratifiedGroupKFold attempts to create folds which preserve the
percentage of samples for each class as much as possible given the
constraint of non-overlapping groups between splits.
Read more in the :ref:`User Guide <cross_validation>`.
Parameters
----------
n_splits : int, default=5
Number of folds. Must be at least 2.
shuffle : bool, default=False
Whether to shuffle each class's samples before splitting into batches.
Note that the samples within each split will not be shuffled.
This implementation can only shuffle groups that have approximately the
same y distribution, no global shuffle will be performed.
random_state : int or RandomState instance, default=None
When `shuffle` is True, `random_state` affects the ordering of the
indices, which controls the randomness of each fold for each class.
Otherwise, leave `random_state` as `None`.
Pass an int for reproducible output across multiple function calls.
See :term:`Glossary <random_state>`.
Examples
--------
>>> import numpy as np
>>> from sklearn.model_selection import StratifiedGroupKFold
>>> X = np.ones((17, 2))
>>> y = np.array([0, 0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0])
>>> groups = np.array([1, 1, 2, 2, 3, 3, 3, 4, 5, 5, 5, 5, 6, 6, 7, 8, 8])
>>> sgkf = StratifiedGroupKFold(n_splits=3)
>>> sgkf.get_n_splits(X, y)
3
>>> print(sgkf)
StratifiedGroupKFold(n_splits=3, random_state=None, shuffle=False)
>>> for i, (train_index, test_index) in enumerate(sgkf.split(X, y, groups)):
... print(f"Fold {i}:")
... print(f" Train: index={train_index}")
... print(f" group={groups[train_index]}")
... print(f" Test: index={test_index}")
... print(f" group={groups[test_index]}")
Fold 0:
Train: index=[ 0 1 2 3 7 8 9 10 11 15 16]
group=[1 1 2 2 4 5 5 5 5 8 8]
Test: index=[ 4 5 6 12 13 14]
group=[3 3 3 6 6 7]
Fold 1:
Train: index=[ 4 5 6 7 8 9 10 11 12 13 14]
group=[3 3 3 4 5 5 5 5 6 6 7]
Test: index=[ 0 1 2 3 15 16]
group=[1 1 2 2 8 8]
Fold 2:
Train: index=[ 0 1 2 3 4 5 6 12 13 14 15 16]
group=[1 1 2 2 3 3 3 6 6 7 8 8]
Test: index=[ 7 8 9 10 11]
group=[4 5 5 5 5]
Notes
-----
The implementation is designed to:
* Mimic the behavior of StratifiedKFold as much as possible for trivial
groups (e.g. when each group contains only one sample).
* Be invariant to class label: relabelling ``y = ["Happy", "Sad"]`` to
``y = [1, 0]`` should not change the indices generated.
* Stratify based on samples as much as possible while keeping
non-overlapping groups constraint. That means that in some cases when
there is a small number of groups containing a large number of samples
the stratification will not be possible and the behavior will be close
to GroupKFold.
See also
--------
StratifiedKFold: Takes class information into account to build folds which
retain class distributions (for binary or multiclass classification
tasks).
GroupKFold: K-fold iterator variant with non-overlapping groups.
"""
def __init__(self, n_splits=5, shuffle=False, random_state=None):
super().__init__(n_splits=n_splits, shuffle=shuffle, random_state=random_state)
def _iter_test_indices(self, X, y, groups):
# Implementation is based on this kaggle kernel:
# https://www.kaggle.com/jakubwasikowski/stratified-group-k-fold-cross-validation
# and is a subject to Apache 2.0 License. You may obtain a copy of the
# License at http://www.apache.org/licenses/LICENSE-2.0
# Changelist:
# - Refactored function to a class following scikit-learn KFold
# interface.
# - Added heuristic for assigning group to the least populated fold in
# cases when all other criteria are equal
# - Swtch from using python ``Counter`` to ``np.unique`` to get class
# distribution
# - Added scikit-learn checks for input: checking that target is binary
# or multiclass, checking passed random state, checking that number
# of splits is less than number of members in each class, checking
# that least populated class has more members than there are splits.
rng = check_random_state(self.random_state)
y = np.asarray(y)
type_of_target_y = type_of_target(y)
allowed_target_types = ("binary", "multiclass")
if type_of_target_y not in allowed_target_types:
raise ValueError(
"Supported target types are: {}. Got {!r} instead.".format(
allowed_target_types, type_of_target_y
)
)
y = column_or_1d(y)
_, y_inv, y_cnt = np.unique(y, return_inverse=True, return_counts=True)
if np.all(self.n_splits > y_cnt):
raise ValueError(
"n_splits=%d cannot be greater than the"
" number of members in each class." % (self.n_splits)
)
n_smallest_class = np.min(y_cnt)
if self.n_splits > n_smallest_class:
warnings.warn(
"The least populated class in y has only %d"
" members, which is less than n_splits=%d."
% (n_smallest_class, self.n_splits),
UserWarning,
)
n_classes = len(y_cnt)
_, groups_inv, groups_cnt = np.unique(
groups, return_inverse=True, return_counts=True
)
y_counts_per_group = np.zeros((len(groups_cnt), n_classes))
for class_idx, group_idx in zip(y_inv, groups_inv):
y_counts_per_group[group_idx, class_idx] += 1
y_counts_per_fold = np.zeros((self.n_splits, n_classes))
groups_per_fold = defaultdict(set)
if self.shuffle:
rng.shuffle(y_counts_per_group)
# Stable sort to keep shuffled order for groups with the same
# class distribution variance
sorted_groups_idx = np.argsort(
-np.std(y_counts_per_group, axis=1), kind="mergesort"
)
for group_idx in sorted_groups_idx:
group_y_counts = y_counts_per_group[group_idx]
best_fold = self._find_best_fold(
y_counts_per_fold=y_counts_per_fold,
y_cnt=y_cnt,
group_y_counts=group_y_counts,
)
y_counts_per_fold[best_fold] += group_y_counts
groups_per_fold[best_fold].add(group_idx)
for i in range(self.n_splits):
test_indices = [
idx
for idx, group_idx in enumerate(groups_inv)
if group_idx in groups_per_fold[i]
]
yield test_indices
def _find_best_fold(self, y_counts_per_fold, y_cnt, group_y_counts):
best_fold = None
min_eval = np.inf
min_samples_in_fold = np.inf
for i in range(self.n_splits):
y_counts_per_fold[i] += group_y_counts
# Summarise the distribution over classes in each proposed fold
std_per_class = np.std(y_counts_per_fold / y_cnt.reshape(1, -1), axis=0)
y_counts_per_fold[i] -= group_y_counts
fold_eval = np.mean(std_per_class)
samples_in_fold = np.sum(y_counts_per_fold[i])
is_current_fold_better = (
fold_eval < min_eval
or np.isclose(fold_eval, min_eval)
and samples_in_fold < min_samples_in_fold
)
if is_current_fold_better:
min_eval = fold_eval
min_samples_in_fold = samples_in_fold
best_fold = i
return best_fold
class TimeSeriesSplit(_BaseKFold):
"""Time Series cross-validator
Provides train/test indices to split time series data samples
that are observed at fixed time intervals, in train/test sets.
In each split, test indices must be higher than before, and thus shuffling
in cross validator is inappropriate.
This cross-validation object is a variation of :class:`KFold`.
In the kth split, it returns first k folds as train set and the
(k+1)th fold as test set.
Note that unlike standard cross-validation methods, successive
training sets are supersets of those that come before them.
Read more in the :ref:`User Guide <time_series_split>`.
.. versionadded:: 0.18
Parameters
----------
n_splits : int, default=5
Number of splits. Must be at least 2.