Mercurial > repos > bgruening > sklearn_data_preprocess
annotate feature_selectors.py @ 24:9e43ee712723 draft
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
author | bgruening |
---|---|
date | Tue, 14 May 2019 18:19:35 -0400 |
parents | |
children |
rev | line source |
---|---|
24
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
1 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
2 DyRFE |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
3 DyRFECV |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
4 MyPipeline |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
5 MyimbPipeline |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
6 check_feature_importances |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
7 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
8 import numpy as np |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
9 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
10 from imblearn import under_sampling, over_sampling, combine |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
11 from imblearn.pipeline import Pipeline as imbPipeline |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
12 from sklearn import (cluster, compose, decomposition, ensemble, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
13 feature_extraction, feature_selection, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
14 gaussian_process, kernel_approximation, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
15 metrics, model_selection, naive_bayes, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
16 neighbors, pipeline, preprocessing, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
17 svm, linear_model, tree, discriminant_analysis) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
18 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
19 from sklearn.base import BaseEstimator |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
20 from sklearn.base import MetaEstimatorMixin, clone, is_classifier |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
21 from sklearn.feature_selection.rfe import _rfe_single_fit, RFE, RFECV |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
22 from sklearn.model_selection import check_cv |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
23 from sklearn.metrics.scorer import check_scoring |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
24 from sklearn.utils import check_X_y, safe_indexing, safe_sqr |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
25 from sklearn.utils._joblib import Parallel, delayed, effective_n_jobs |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
26 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
27 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
28 class DyRFE(RFE): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
29 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
30 Mainly used with DyRFECV |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
31 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
32 Parameters |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
33 ---------- |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
34 estimator : object |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
35 A supervised learning estimator with a ``fit`` method that provides |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
36 information about feature importance either through a ``coef_`` |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
37 attribute or through a ``feature_importances_`` attribute. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
38 n_features_to_select : int or None (default=None) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
39 The number of features to select. If `None`, half of the features |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
40 are selected. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
41 step : int, float or list, optional (default=1) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
42 If greater than or equal to 1, then ``step`` corresponds to the |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
43 (integer) number of features to remove at each iteration. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
44 If within (0.0, 1.0), then ``step`` corresponds to the percentage |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
45 (rounded down) of features to remove at each iteration. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
46 If list, a series of steps of features to remove at each iteration. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
47 Iterations stops when steps finish |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
48 verbose : int, (default=0) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
49 Controls verbosity of output. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
50 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
51 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
52 def __init__(self, estimator, n_features_to_select=None, step=1, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
53 verbose=0): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
54 super(DyRFE, self).__init__(estimator, n_features_to_select, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
55 step, verbose) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
56 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
57 def _fit(self, X, y, step_score=None): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
58 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
59 if type(self.step) is not list: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
60 return super(DyRFE, self)._fit(X, y, step_score) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
61 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
62 # dynamic step |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
63 X, y = check_X_y(X, y, "csc") |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
64 # Initialization |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
65 n_features = X.shape[1] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
66 if self.n_features_to_select is None: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
67 n_features_to_select = n_features // 2 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
68 else: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
69 n_features_to_select = self.n_features_to_select |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
70 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
71 step = [] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
72 for s in self.step: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
73 if 0.0 < s < 1.0: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
74 step.append(int(max(1, s * n_features))) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
75 else: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
76 step.append(int(s)) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
77 if s <= 0: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
78 raise ValueError("Step must be >0") |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
79 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
80 support_ = np.ones(n_features, dtype=np.bool) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
81 ranking_ = np.ones(n_features, dtype=np.int) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
82 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
83 if step_score: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
84 self.scores_ = [] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
85 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
86 step_i = 0 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
87 # Elimination |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
88 while np.sum(support_) > n_features_to_select and step_i < len(step): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
89 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
90 # if last step is 1, will keep loop |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
91 if step_i == len(step) - 1 and step[step_i] != 0: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
92 step.append(step[step_i]) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
93 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
94 # Remaining features |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
95 features = np.arange(n_features)[support_] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
96 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
97 # Rank the remaining features |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
98 estimator = clone(self.estimator) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
99 if self.verbose > 0: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
100 print("Fitting estimator with %d features." % np.sum(support_)) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
101 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
102 estimator.fit(X[:, features], y) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
103 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
104 # Get coefs |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
105 if hasattr(estimator, 'coef_'): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
106 coefs = estimator.coef_ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
107 else: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
108 coefs = getattr(estimator, 'feature_importances_', None) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
109 if coefs is None: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
110 raise RuntimeError('The classifier does not expose ' |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
111 '"coef_" or "feature_importances_" ' |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
112 'attributes') |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
113 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
114 # Get ranks |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
115 if coefs.ndim > 1: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
116 ranks = np.argsort(safe_sqr(coefs).sum(axis=0)) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
117 else: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
118 ranks = np.argsort(safe_sqr(coefs)) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
119 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
120 # for sparse case ranks is matrix |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
121 ranks = np.ravel(ranks) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
122 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
123 # Eliminate the worse features |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
124 threshold =\ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
125 min(step[step_i], np.sum(support_) - n_features_to_select) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
126 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
127 # Compute step score on the previous selection iteration |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
128 # because 'estimator' must use features |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
129 # that have not been eliminated yet |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
130 if step_score: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
131 self.scores_.append(step_score(estimator, features)) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
132 support_[features[ranks][:threshold]] = False |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
133 ranking_[np.logical_not(support_)] += 1 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
134 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
135 step_i += 1 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
136 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
137 # Set final attributes |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
138 features = np.arange(n_features)[support_] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
139 self.estimator_ = clone(self.estimator) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
140 self.estimator_.fit(X[:, features], y) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
141 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
142 # Compute step score when only n_features_to_select features left |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
143 if step_score: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
144 self.scores_.append(step_score(self.estimator_, features)) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
145 self.n_features_ = support_.sum() |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
146 self.support_ = support_ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
147 self.ranking_ = ranking_ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
148 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
149 return self |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
150 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
151 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
152 class DyRFECV(RFECV, MetaEstimatorMixin): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
153 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
154 Compared with RFECV, DyRFECV offers flexiable `step` to eleminate |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
155 features, in the format of list, while RFECV supports only fixed number |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
156 of `step`. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
157 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
158 Parameters |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
159 ---------- |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
160 estimator : object |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
161 A supervised learning estimator with a ``fit`` method that provides |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
162 information about feature importance either through a ``coef_`` |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
163 attribute or through a ``feature_importances_`` attribute. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
164 step : int or float, optional (default=1) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
165 If greater than or equal to 1, then ``step`` corresponds to the |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
166 (integer) number of features to remove at each iteration. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
167 If within (0.0, 1.0), then ``step`` corresponds to the percentage |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
168 (rounded down) of features to remove at each iteration. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
169 If list, a series of step to remove at each iteration. iteration stopes |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
170 when finishing all steps |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
171 Note that the last iteration may remove fewer than ``step`` features in |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
172 order to reach ``min_features_to_select``. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
173 min_features_to_select : int, (default=1) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
174 The minimum number of features to be selected. This number of features |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
175 will always be scored, even if the difference between the original |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
176 feature count and ``min_features_to_select`` isn't divisible by |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
177 ``step``. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
178 cv : int, cross-validation generator or an iterable, optional |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
179 Determines the cross-validation splitting strategy. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
180 Possible inputs for cv are: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
181 - None, to use the default 3-fold cross-validation, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
182 - integer, to specify the number of folds. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
183 - :term:`CV splitter`, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
184 - An iterable yielding (train, test) splits as arrays of indices. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
185 For integer/None inputs, if ``y`` is binary or multiclass, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
186 :class:`sklearn.model_selection.StratifiedKFold` is used. If the |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
187 estimator is a classifier or if ``y`` is neither binary nor multiclass, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
188 :class:`sklearn.model_selection.KFold` is used. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
189 Refer :ref:`User Guide <cross_validation>` for the various |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
190 cross-validation strategies that can be used here. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
191 .. versionchanged:: 0.20 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
192 ``cv`` default value of None will change from 3-fold to 5-fold |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
193 in v0.22. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
194 scoring : string, callable or None, optional, (default=None) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
195 A string (see model evaluation documentation) or |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
196 a scorer callable object / function with signature |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
197 ``scorer(estimator, X, y)``. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
198 verbose : int, (default=0) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
199 Controls verbosity of output. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
200 n_jobs : int or None, optional (default=None) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
201 Number of cores to run in parallel while fitting across folds. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
202 ``None`` means 1 unless in a :obj:`joblib.parallel_backend` context. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
203 ``-1`` means using all processors. See :term:`Glossary <n_jobs>` |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
204 for more details. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
205 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
206 def __init__(self, estimator, step=1, min_features_to_select=1, cv='warn', |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
207 scoring=None, verbose=0, n_jobs=None): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
208 super(DyRFECV, self).__init__( |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
209 estimator, step=step, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
210 min_features_to_select=min_features_to_select, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
211 cv=cv, scoring=scoring, verbose=verbose, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
212 n_jobs=n_jobs) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
213 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
214 def fit(self, X, y, groups=None): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
215 """Fit the RFE model and automatically tune the number of selected |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
216 features. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
217 Parameters |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
218 ---------- |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
219 X : {array-like, sparse matrix}, shape = [n_samples, n_features] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
220 Training vector, where `n_samples` is the number of samples and |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
221 `n_features` is the total number of features. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
222 y : array-like, shape = [n_samples] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
223 Target values (integers for classification, real numbers for |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
224 regression). |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
225 groups : array-like, shape = [n_samples], optional |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
226 Group labels for the samples used while splitting the dataset into |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
227 train/test set. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
228 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
229 if type(self.step) is not list: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
230 return super(DyRFECV, self).fit(X, y, groups) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
231 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
232 X, y = check_X_y(X, y, "csr") |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
233 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
234 # Initialization |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
235 cv = check_cv(self.cv, y, is_classifier(self.estimator)) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
236 scorer = check_scoring(self.estimator, scoring=self.scoring) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
237 n_features = X.shape[1] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
238 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
239 step = [] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
240 for s in self.step: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
241 if 0.0 < s < 1.0: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
242 step.append(int(max(1, s * n_features))) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
243 else: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
244 step.append(int(s)) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
245 if s <= 0: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
246 raise ValueError("Step must be >0") |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
247 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
248 # Build an RFE object, which will evaluate and score each possible |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
249 # feature count, down to self.min_features_to_select |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
250 rfe = DyRFE(estimator=self.estimator, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
251 n_features_to_select=self.min_features_to_select, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
252 step=self.step, verbose=self.verbose) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
253 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
254 # Determine the number of subsets of features by fitting across |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
255 # the train folds and choosing the "features_to_select" parameter |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
256 # that gives the least averaged error across all folds. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
257 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
258 # Note that joblib raises a non-picklable error for bound methods |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
259 # even if n_jobs is set to 1 with the default multiprocessing |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
260 # backend. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
261 # This branching is done so that to |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
262 # make sure that user code that sets n_jobs to 1 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
263 # and provides bound methods as scorers is not broken with the |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
264 # addition of n_jobs parameter in version 0.18. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
265 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
266 if effective_n_jobs(self.n_jobs) == 1: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
267 parallel, func = list, _rfe_single_fit |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
268 else: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
269 parallel = Parallel(n_jobs=self.n_jobs) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
270 func = delayed(_rfe_single_fit) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
271 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
272 scores = parallel( |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
273 func(rfe, self.estimator, X, y, train, test, scorer) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
274 for train, test in cv.split(X, y, groups)) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
275 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
276 scores = np.sum(scores, axis=0) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
277 diff = int(scores.shape[0]) - len(step) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
278 if diff > 0: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
279 step = np.r_[step, [step[-1]] * diff] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
280 scores_rev = scores[::-1] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
281 argmax_idx = len(scores) - np.argmax(scores_rev) - 1 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
282 n_features_to_select = max( |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
283 n_features - sum(step[:argmax_idx]), |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
284 self.min_features_to_select) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
285 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
286 # Re-execute an elimination with best_k over the whole set |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
287 rfe = DyRFE(estimator=self.estimator, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
288 n_features_to_select=n_features_to_select, step=self.step, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
289 verbose=self.verbose) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
290 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
291 rfe.fit(X, y) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
292 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
293 # Set final attributes |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
294 self.support_ = rfe.support_ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
295 self.n_features_ = rfe.n_features_ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
296 self.ranking_ = rfe.ranking_ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
297 self.estimator_ = clone(self.estimator) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
298 self.estimator_.fit(self.transform(X), y) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
299 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
300 # Fixing a normalization error, n is equal to get_n_splits(X, y) - 1 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
301 # here, the scores are normalized by get_n_splits(X, y) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
302 self.grid_scores_ = scores[::-1] / cv.get_n_splits(X, y, groups) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
303 return self |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
304 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
305 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
306 class MyPipeline(pipeline.Pipeline): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
307 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
308 Extend pipeline object to have feature_importances_ attribute |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
309 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
310 def fit(self, X, y=None, **fit_params): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
311 super(MyPipeline, self).fit(X, y, **fit_params) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
312 estimator = self.steps[-1][-1] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
313 if hasattr(estimator, 'coef_'): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
314 coefs = estimator.coef_ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
315 else: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
316 coefs = getattr(estimator, 'feature_importances_', None) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
317 if coefs is None: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
318 raise RuntimeError('The estimator in the pipeline does not expose ' |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
319 '"coef_" or "feature_importances_" ' |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
320 'attributes') |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
321 self.feature_importances_ = coefs |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
322 return self |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
323 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
324 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
325 class MyimbPipeline(imbPipeline): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
326 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
327 Extend imblance pipeline object to have feature_importances_ attribute |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
328 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
329 def fit(self, X, y=None, **fit_params): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
330 super(MyimbPipeline, self).fit(X, y, **fit_params) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
331 estimator = self.steps[-1][-1] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
332 if hasattr(estimator, 'coef_'): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
333 coefs = estimator.coef_ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
334 else: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
335 coefs = getattr(estimator, 'feature_importances_', None) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
336 if coefs is None: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
337 raise RuntimeError('The estimator in the pipeline does not expose ' |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
338 '"coef_" or "feature_importances_" ' |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
339 'attributes') |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
340 self.feature_importances_ = coefs |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
341 return self |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
342 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
343 |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
344 def check_feature_importances(estimator): |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
345 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
346 For pipeline object which has no feature_importances_ property, |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
347 this function returns the same comfigured pipeline object with |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
348 attached the last estimator's feature_importances_. |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
349 """ |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
350 if estimator.__class__.__module__ == 'sklearn.pipeline': |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
351 pipeline_steps = estimator.get_params()['steps'] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
352 estimator = MyPipeline(pipeline_steps) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
353 elif estimator.__class__.__module__ == 'imblearn.pipeline': |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
354 pipeline_steps = estimator.get_params()['steps'] |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
355 estimator = MyimbPipeline(pipeline_steps) |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
356 else: |
9e43ee712723
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/sklearn commit c0a3a186966888e5787335a7628bf0a4382637e7
bgruening
parents:
diff
changeset
|
357 return estimator |