Mercurial > repos > bgruening > sklearn_model_validation

annotate model_validations.py @ 18:492d34a75de6 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit 49522db5f2dc8a571af49e3f38e80c22571068f4
author bgruening
date Tue, 09 Jul 2019 19:39:58 -0400
parents cf9aa11b91c8
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
17
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
1 """
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
2 class
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
3 -----
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
4 OrderedKFold
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
5 RepeatedOrderedKold
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
6
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
7
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
8 function
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
9 --------
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
10 train_test_split
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
11 """
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
12
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
13 import numpy as np
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
14 import warnings
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
15
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
16 from itertools import chain
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
17 from math import ceil, floor
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
18 from sklearn.model_selection import (GroupShuffleSplit, ShuffleSplit,
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
19 StratifiedShuffleSplit)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
20 from sklearn.model_selection._split import _BaseKFold, _RepeatedSplits
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
21 from sklearn.utils import check_random_state, indexable, safe_indexing
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
22 from sklearn.utils.validation import _num_samples, check_array
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
23
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
24
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
25 def _validate_shuffle_split(n_samples, test_size, train_size,
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
26 default_test_size=None):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
27 """
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
28 Validation helper to check if the test/test sizes are meaningful wrt to the
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
29 size of the data (n_samples)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
30 """
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
31 if test_size is None and train_size is None:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
32 test_size = default_test_size
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
33
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
34 test_size_type = np.asarray(test_size).dtype.kind
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
35 train_size_type = np.asarray(train_size).dtype.kind
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
36
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
37 if (test_size_type == 'i' and (test_size >= n_samples or test_size <= 0)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
38 or test_size_type == 'f' and (test_size <= 0 or test_size >= 1)):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
39 raise ValueError('test_size={0} should be either positive and smaller'
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
40 ' than the number of samples {1} or a float in the '
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
41 '(0, 1) range'.format(test_size, n_samples))
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
42
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
43 if (train_size_type == 'i' and (train_size >= n_samples or train_size <= 0)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
44 or train_size_type == 'f' and (train_size <= 0 or train_size >= 1)):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
45 raise ValueError('train_size={0} should be either positive and smaller'
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
46 ' than the number of samples {1} or a float in the '
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
47 '(0, 1) range'.format(train_size, n_samples))
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
48
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
49 if train_size is not None and train_size_type not in ('i', 'f'):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
50 raise ValueError("Invalid value for train_size: {}".format(train_size))
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
51 if test_size is not None and test_size_type not in ('i', 'f'):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
52 raise ValueError("Invalid value for test_size: {}".format(test_size))
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
53
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
54 if (train_size_type == 'f' and test_size_type == 'f' and
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
55 train_size + test_size > 1):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
56 raise ValueError(
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
57 'The sum of test_size and train_size = {}, should be in the (0, 1)'
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
58 ' range. Reduce test_size and/or train_size.'
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
59 .format(train_size + test_size))
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
60
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
61 if test_size_type == 'f':
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
62 n_test = ceil(test_size * n_samples)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
63 elif test_size_type == 'i':
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
64 n_test = float(test_size)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
65
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
66 if train_size_type == 'f':
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
67 n_train = floor(train_size * n_samples)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
68 elif train_size_type == 'i':
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
69 n_train = float(train_size)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
70
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
71 if train_size is None:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
72 n_train = n_samples - n_test
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
73 elif test_size is None:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
74 n_test = n_samples - n_train
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
75
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
76 if n_train + n_test > n_samples:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
77 raise ValueError('The sum of train_size and test_size = %d, '
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
78 'should be smaller than the number of '
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
79 'samples %d. Reduce test_size and/or '
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
80 'train_size.' % (n_train + n_test, n_samples))
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
81
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
82 n_train, n_test = int(n_train), int(n_test)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
83
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
84 if n_train == 0:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
85 raise ValueError(
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
86 'With n_samples={}, test_size={} and train_size={}, the '
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
87 'resulting train set will be empty. Adjust any of the '
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
88 'aforementioned parameters.'.format(n_samples, test_size,
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
89 train_size)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
90 )
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
91
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
92 return n_train, n_test
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
93
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
94
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
95 def train_test_split(*arrays, **options):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
96 """Extend sklearn.model_selection.train_test_slit to have group split.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
97
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
98 Parameters
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
99 ----------
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
100 *arrays : sequence of indexables with same length / shape[0]
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
101 Allowed inputs are lists, numpy arrays, scipy-sparse
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
102 matrices or pandas dataframes.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
103
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
104 test_size : float, int or None, optional (default=None)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
105 If float, should be between 0.0 and 1.0 and represent the proportion
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
106 of the dataset to include in the test split. If int, represents the
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
107 absolute number of test samples. If None, the value is set to the
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
108 complement of the train size. If ``train_size`` is also None, it will
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
109 be set to 0.25.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
110
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
111 train_size : float, int, or None, (default=None)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
112 If float, should be between 0.0 and 1.0 and represent the
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
113 proportion of the dataset to include in the train split. If
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
114 int, represents the absolute number of train samples. If None,
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
115 the value is automatically set to the complement of the test size.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
116
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
117 random_state : int, RandomState instance or None, optional (default=None)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
118 If int, random_state is the seed used by the random number generator;
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
119 If RandomState instance, random_state is the random number generator;
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
120 If None, the random number generator is the RandomState instance used
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
121 by `np.random`.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
122
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
123 shuffle : None or str (default='simple')
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
124 How to shuffle the data before splitting.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
125 None, no shuffle.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
126 For str, one of 'simple', 'stratified' and 'group', corresponding to
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
127 `ShuffleSplit`, `StratifiedShuffleSplit` and `GroupShuffleSplit`,
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
128 respectively.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
129
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
130 labels : array-like or None (default=None)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
131 Ignored if shuffle is None or 'simple'.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
132 When shuffle='stratified', this array is used as class labels.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
133 When shuffle='group', this array is used as groups.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
134
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
135 Returns
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
136 -------
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
137 splitting : list, length=2 * len(arrays)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
138 List containing train-test split of inputs.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
139
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
140 """
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
141 n_arrays = len(arrays)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
142 if n_arrays == 0:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
143 raise ValueError("At least one array required as input")
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
144 test_size = options.pop('test_size', None)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
145 train_size = options.pop('train_size', None)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
146 random_state = options.pop('random_state', None)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
147 shuffle = options.pop('shuffle', 'simple')
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
148 labels = options.pop('labels', None)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
149
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
150 if options:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
151 raise TypeError("Invalid parameters passed: %s" % str(options))
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
152
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
153 arrays = indexable(*arrays)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
154
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
155 n_samples = _num_samples(arrays[0])
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
156 if shuffle == 'group':
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
157 if labels is None:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
158 raise ValueError("When shuffle='group', "
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
159 "labels should not be None!")
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
160 labels = check_array(labels, ensure_2d=False, dtype=None)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
161 uniques = np.unique(labels)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
162 n_samples = uniques.size
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
163
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
164 n_train, n_test = _validate_shuffle_split(n_samples, test_size, train_size,
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
165 default_test_size=0.25)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
166
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
167 shuffle_options = dict(test_size=n_test,
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
168 train_size=n_train,
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
169 random_state=random_state)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
170
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
171 if shuffle is None:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
172 if labels is not None:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
173 warnings.warn("The `labels` is ignored for "
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
174 "shuffle being None!")
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
175
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
176 train = np.arange(n_train)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
177 test = np.arange(n_train, n_train + n_test)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
178
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
179 elif shuffle == 'simple':
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
180 if labels is not None:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
181 warnings.warn("The `labels` is not needed and therefore "
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
182 "ignored for ShuffleSplit, as shuffle='simple'!")
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
183
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
184 cv = ShuffleSplit(**shuffle_options)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
185 train, test = next(cv.split(X=arrays[0], y=None))
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
186
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
187 elif shuffle == 'stratified':
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
188 cv = StratifiedShuffleSplit(**shuffle_options)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
189 train, test = next(cv.split(X=arrays[0], y=labels))
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
190
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
191 elif shuffle == 'group':
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
192 cv = GroupShuffleSplit(**shuffle_options)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
193 train, test = next(cv.split(X=arrays[0], y=None, groups=labels))
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
194
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
195 else:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
196 raise ValueError("The argument `shuffle` only supports None, "
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
197 "'simple', 'stratified' and 'group', but got `%s`!"
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
198 % shuffle)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
199
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
200 return list(chain.from_iterable((safe_indexing(a, train),
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
201 safe_indexing(a, test)) for a in arrays))
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
202
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
203
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
204 class OrderedKFold(_BaseKFold):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
205 """
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
206 Split into K fold based on ordered target value
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
207
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
208 Parameters
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
209 ----------
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
210 n_splits : int, default=3
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
211 Number of folds. Must be at least 2.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
212 shuffle: bool
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
213 random_state: None or int
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
214 """
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
215
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
216 def __init__(self, n_splits=3, shuffle=False, random_state=None):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
217 super(OrderedKFold, self).__init__(n_splits, shuffle, random_state)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
218
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
219 def _iter_test_indices(self, X, y, groups=None):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
220 n_samples = _num_samples(X)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
221 n_splits = self.n_splits
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
222 y = np.asarray(y)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
223 sorted_index = np.argsort(y)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
224 if self.shuffle:
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
225 current = 0
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
226 rng = check_random_state(self.random_state)
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
227 for i in range(n_samples // int(n_splits)):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
228 start, stop = current, current + n_splits
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
229 rng.shuffle(sorted_index[start:stop])
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
230 current = stop
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
231 rng.shuffle(sorted_index[current:])
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
232
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
233 for i in range(n_splits):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
234 yield sorted_index[i:n_samples:n_splits]
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
235
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
236
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
237 class RepeatedOrderedKFold(_RepeatedSplits):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
238 """ Repeated OrderedKFold runs mutiple times with different randomization.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
239
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
240 Parameters
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
241 ----------
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
242 n_splits : int, default=5
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
243 Number of folds. Must be at least 2.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
244
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
245 n_repeats : int, default=5
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
246 Number of times cross-validator to be repeated.
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
247
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
248 random_state: int, RandomState instance or None. Optional
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
249 """
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
250 def __init__(self, n_splits=5, n_repeats=5, random_state=None):
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
251 super(RepeatedOrderedKFold, self).__init__(
cf9aa11b91c8 planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit ab963ec9498bd05d2fb2f24f75adb2fccae7958c
bgruening
parents:
diff changeset
252 OrderedKFold, n_repeats, random_state, n_splits=n_splits)