forked from scikit-learn/scikit-learn.github.io
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcompose.html
838 lines (801 loc) · 86.8 KB
/
compose.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
<meta charset="utf-8">
<meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
<meta property="og:title" content="6.1. Pipelines and composite estimators" />
<meta property="og:type" content="website" />
<meta property="og:url" content="https://scikit-learn/stable/modules/compose.html" />
<meta property="og:site_name" content="scikit-learn" />
<meta property="og:description" content="Transformers are usually combined with classifiers, regressors or other estimators to build a composite estimator. The most common tool is a Pipeline. Pipeline is often used in combination with Fea..." />
<meta property="og:image" content="https://scikit-learn.org/stable/_static/scikit-learn-logo-small.png" />
<meta property="og:image:alt" content="scikit-learn" />
<meta name="description" content="Transformers are usually combined with classifiers, regressors or other estimators to build a composite estimator. The most common tool is a Pipeline. Pipeline is often used in combination with Fea..." />
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>6.1. Pipelines and composite estimators — scikit-learn 1.4.dev0 documentation</title>
<link rel="canonical" href="http://scikit-learn.org/stable/modules/compose.html" />
<link rel="shortcut icon" href="../_static/favicon.ico"/>
<link rel="stylesheet" href="../_static/css/vendor/bootstrap.min.css" type="text/css" />
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
<link rel="stylesheet" href="../_static/copybutton.css" type="text/css" />
<link rel="stylesheet" href="../_static/plot_directive.css" type="text/css" />
<link rel="stylesheet" href="https://fonts.googleapis.com/css?family=Vibur" type="text/css" />
<link rel="stylesheet" href="../_static/jupyterlite_sphinx.css" type="text/css" />
<link rel="stylesheet" href="../_static/sg_gallery.css" type="text/css" />
<link rel="stylesheet" href="../_static/sg_gallery-binder.css" type="text/css" />
<link rel="stylesheet" href="../_static/sg_gallery-dataframe.css" type="text/css" />
<link rel="stylesheet" href="../_static/sg_gallery-rendered-html.css" type="text/css" />
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
<script id="documentation_options" data-url_root="../" src="../_static/documentation_options.js"></script>
<script src="../_static/js/vendor/jquery-3.6.3.slim.min.js"></script>
</head>
<body>
<nav id="navbar" class="sk-docs-navbar navbar navbar-expand-md navbar-light bg-light py-0">
<div class="container-fluid sk-docs-container px-0">
<a class="navbar-brand py-0" href="../index.html">
<img
class="sk-brand-img"
src="../_static/scikit-learn-logo-small.png"
alt="logo"/>
</a>
<button
id="sk-navbar-toggler"
class="navbar-toggler"
type="button"
data-toggle="collapse"
data-target="#navbarSupportedContent"
aria-controls="navbarSupportedContent"
aria-expanded="false"
aria-label="Toggle navigation"
>
<span class="navbar-toggler-icon"></span>
</button>
<div class="sk-navbar-collapse collapse navbar-collapse" id="navbarSupportedContent">
<ul class="navbar-nav mr-auto">
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../install.html">Install</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../user_guide.html">User Guide</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" href="classes.html">API</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" href="../auto_examples/index.html">Examples</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link" target="_blank" rel="noopener noreferrer" href="https://blog.scikit-learn.org/">Community</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../getting_started.html" >Getting Started</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../tutorial/index.html" >Tutorial</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../whats_new/v1.4.html" >What's new</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../glossary.html" >Glossary</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../developers/index.html" >Development</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../faq.html" >FAQ</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../support.html" >Support</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../related_projects.html" >Related packages</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../roadmap.html" >Roadmap</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../governance.html" >Governance</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="../about.html" >About us</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="https://github.com/scikit-learn/scikit-learn" >GitHub</a>
</li>
<li class="nav-item">
<a class="sk-nav-link nav-link nav-more-item-mobile-items" href="https://scikit-learn.org/dev/versions.html" >Other Versions and Download</a>
</li>
<li class="nav-item dropdown nav-more-item-dropdown">
<a class="sk-nav-link nav-link dropdown-toggle" href="#" id="navbarDropdown" role="button" data-toggle="dropdown" aria-haspopup="true" aria-expanded="false">More</a>
<div class="dropdown-menu" aria-labelledby="navbarDropdown">
<a class="sk-nav-dropdown-item dropdown-item" href="../getting_started.html" >Getting Started</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../tutorial/index.html" >Tutorial</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../whats_new/v1.4.html" >What's new</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../glossary.html" >Glossary</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../developers/index.html" >Development</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../faq.html" >FAQ</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../support.html" >Support</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../related_projects.html" >Related packages</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../roadmap.html" >Roadmap</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../governance.html" >Governance</a>
<a class="sk-nav-dropdown-item dropdown-item" href="../about.html" >About us</a>
<a class="sk-nav-dropdown-item dropdown-item" href="https://github.com/scikit-learn/scikit-learn" >GitHub</a>
<a class="sk-nav-dropdown-item dropdown-item" href="https://scikit-learn.org/dev/versions.html" >Other Versions and Download</a>
</div>
</li>
</ul>
<div id="searchbox" role="search">
<div class="searchformwrapper">
<form class="search" action="../search.html" method="get">
<input class="sk-search-text-input" type="text" name="q" aria-labelledby="searchlabel" />
<input class="sk-search-text-btn" type="submit" value="Go" />
</form>
</div>
</div>
</div>
</div>
</nav>
<div class="d-flex" id="sk-doc-wrapper">
<input type="checkbox" name="sk-toggle-checkbox" id="sk-toggle-checkbox">
<label id="sk-sidemenu-toggle" class="sk-btn-toggle-toc btn sk-btn-primary" for="sk-toggle-checkbox">Toggle Menu</label>
<div id="sk-sidebar-wrapper" class="border-right">
<div class="sk-sidebar-toc-wrapper">
<div class="btn-group w-100 mb-2" role="group" aria-label="rellinks">
<a href="../data_transforms.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="6. Dataset transformations">Prev</a><a href="../data_transforms.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="6. Dataset transformations">Up</a>
<a href="feature_extraction.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="6.2. Feature extraction">Next</a>
</div>
<div class="alert alert-danger p-1 mb-2" role="alert">
<p class="text-center mb-0">
<strong>scikit-learn 1.4.dev0</strong><br/>
<a href="http://scikit-learn.org/dev/versions.html">Other versions</a>
</p>
</div>
<div class="alert alert-warning p-1 mb-2" role="alert">
<p class="text-center mb-0">
Please <a class="font-weight-bold" href="../about.html#citing-scikit-learn"><string>cite us</string></a> if you use the software.
</p>
</div>
<div class="sk-sidebar-toc">
<ul>
<li><a class="reference internal" href="#">6.1. Pipelines and composite estimators</a><ul>
<li><a class="reference internal" href="#pipeline-chaining-estimators">6.1.1. Pipeline: chaining estimators</a><ul>
<li><a class="reference internal" href="#usage">6.1.1.1. Usage</a><ul>
<li><a class="reference internal" href="#construction">6.1.1.1.1. Construction</a></li>
<li><a class="reference internal" href="#accessing-steps">6.1.1.1.2. Accessing steps</a></li>
<li><a class="reference internal" href="#nested-parameters">6.1.1.1.3. Nested parameters</a></li>
</ul>
</li>
<li><a class="reference internal" href="#notes">6.1.1.2. Notes</a></li>
<li><a class="reference internal" href="#caching-transformers-avoid-repeated-computation">6.1.1.3. Caching transformers: avoid repeated computation</a></li>
</ul>
</li>
<li><a class="reference internal" href="#transforming-target-in-regression">6.1.2. Transforming target in regression</a></li>
<li><a class="reference internal" href="#featureunion-composite-feature-spaces">6.1.3. FeatureUnion: composite feature spaces</a><ul>
<li><a class="reference internal" href="#id1">6.1.3.1. Usage</a></li>
</ul>
</li>
<li><a class="reference internal" href="#columntransformer-for-heterogeneous-data">6.1.4. ColumnTransformer for heterogeneous data</a></li>
<li><a class="reference internal" href="#visualizing-composite-estimators">6.1.5. Visualizing Composite Estimators</a></li>
</ul>
</li>
</ul>
</div>
</div>
</div>
<div id="sk-page-content-wrapper">
<div class="sk-page-content container-fluid body px-md-3" role="main">
<section id="pipelines-and-composite-estimators">
<span id="combining-estimators"></span><h1><span class="section-number">6.1. </span>Pipelines and composite estimators<a class="headerlink" href="#pipelines-and-composite-estimators" title="Permalink to this heading">¶</a></h1>
<p>Transformers are usually combined with classifiers, regressors or other
estimators to build a composite estimator. The most common tool is a
<a class="reference internal" href="#pipeline"><span class="std std-ref">Pipeline</span></a>. Pipeline is often used in combination with
<a class="reference internal" href="#feature-union"><span class="std std-ref">FeatureUnion</span></a> which concatenates the output of
transformers into a composite feature space. <a class="reference internal" href="#transformed-target-regressor"><span class="std std-ref">TransformedTargetRegressor</span></a> deals with transforming the <a class="reference internal" href="../glossary.html#term-target"><span class="xref std std-term">target</span></a>
(i.e. log-transform <a class="reference internal" href="../glossary.html#term-y"><span class="xref std std-term">y</span></a>). In contrast, Pipelines only transform the
observed data (<a class="reference internal" href="../glossary.html#term-X"><span class="xref std std-term">X</span></a>).</p>
<section id="pipeline-chaining-estimators">
<span id="pipeline"></span><h2><span class="section-number">6.1.1. </span>Pipeline: chaining estimators<a class="headerlink" href="#pipeline-chaining-estimators" title="Permalink to this heading">¶</a></h2>
<p><a class="reference internal" href="generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">Pipeline</span></code></a> can be used to chain multiple estimators
into one. This is useful as there is often a fixed sequence
of steps in processing the data, for example feature selection, normalization
and classification. <a class="reference internal" href="generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">Pipeline</span></code></a> serves multiple purposes here:</p>
<dl class="simple">
<dt>Convenience and encapsulation</dt><dd><p>You only have to call <a class="reference internal" href="../glossary.html#term-fit"><span class="xref std std-term">fit</span></a> and <a class="reference internal" href="../glossary.html#term-predict"><span class="xref std std-term">predict</span></a> once on your
data to fit a whole sequence of estimators.</p>
</dd>
<dt>Joint parameter selection</dt><dd><p>You can <a class="reference internal" href="grid_search.html#grid-search"><span class="std std-ref">grid search</span></a>
over parameters of all estimators in the pipeline at once.</p>
</dd>
<dt>Safety</dt><dd><p>Pipelines help avoid leaking statistics from your test data into the
trained model in cross-validation, by ensuring that the same samples are
used to train the transformers and predictors.</p>
</dd>
</dl>
<p>All estimators in a pipeline, except the last one, must be transformers
(i.e. must have a <a class="reference internal" href="../glossary.html#term-transform"><span class="xref std std-term">transform</span></a> method).
The last estimator may be any type (transformer, classifier, etc.).</p>
<section id="usage">
<h3><span class="section-number">6.1.1.1. </span>Usage<a class="headerlink" href="#usage" title="Permalink to this heading">¶</a></h3>
<section id="construction">
<h4><span class="section-number">6.1.1.1.1. </span>Construction<a class="headerlink" href="#construction" title="Permalink to this heading">¶</a></h4>
<p>The <a class="reference internal" href="generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">Pipeline</span></code></a> is built using a list of <code class="docutils literal notranslate"><span class="pre">(key,</span> <span class="pre">value)</span></code> pairs, where
the <code class="docutils literal notranslate"><span class="pre">key</span></code> is a string containing the name you want to give this step and <code class="docutils literal notranslate"><span class="pre">value</span></code>
is an estimator object:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.pipeline</span> <span class="kn">import</span> <span class="n">Pipeline</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.svm</span> <span class="kn">import</span> <span class="n">SVC</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.decomposition</span> <span class="kn">import</span> <span class="n">PCA</span>
<span class="gp">>>> </span><span class="n">estimators</span> <span class="o">=</span> <span class="p">[(</span><span class="s1">'reduce_dim'</span><span class="p">,</span> <span class="n">PCA</span><span class="p">()),</span> <span class="p">(</span><span class="s1">'clf'</span><span class="p">,</span> <span class="n">SVC</span><span class="p">())]</span>
<span class="gp">>>> </span><span class="n">pipe</span> <span class="o">=</span> <span class="n">Pipeline</span><span class="p">(</span><span class="n">estimators</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">pipe</span>
<span class="go">Pipeline(steps=[('reduce_dim', PCA()), ('clf', SVC())])</span>
</pre></div>
</div>
<p>The utility function <a class="reference internal" href="generated/sklearn.pipeline.make_pipeline.html#sklearn.pipeline.make_pipeline" title="sklearn.pipeline.make_pipeline"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_pipeline</span></code></a> is a shorthand
for constructing pipelines;
it takes a variable number of estimators and returns a pipeline,
filling in the names automatically:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.pipeline</span> <span class="kn">import</span> <span class="n">make_pipeline</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.naive_bayes</span> <span class="kn">import</span> <span class="n">MultinomialNB</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.preprocessing</span> <span class="kn">import</span> <span class="n">Binarizer</span>
<span class="gp">>>> </span><span class="n">make_pipeline</span><span class="p">(</span><span class="n">Binarizer</span><span class="p">(),</span> <span class="n">MultinomialNB</span><span class="p">())</span>
<span class="go">Pipeline(steps=[('binarizer', Binarizer()), ('multinomialnb', MultinomialNB())])</span>
</pre></div>
</div>
</section>
<section id="accessing-steps">
<h4><span class="section-number">6.1.1.1.2. </span>Accessing steps<a class="headerlink" href="#accessing-steps" title="Permalink to this heading">¶</a></h4>
<p>The estimators of a pipeline are stored as a list in the <code class="docutils literal notranslate"><span class="pre">steps</span></code> attribute,
but can be accessed by index or name by indexing (with <code class="docutils literal notranslate"><span class="pre">[idx]</span></code>) the
Pipeline:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">pipe</span><span class="o">.</span><span class="n">steps</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
<span class="go">('reduce_dim', PCA())</span>
<span class="gp">>>> </span><span class="n">pipe</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
<span class="go">PCA()</span>
<span class="gp">>>> </span><span class="n">pipe</span><span class="p">[</span><span class="s1">'reduce_dim'</span><span class="p">]</span>
<span class="go">PCA()</span>
</pre></div>
</div>
<p>Pipeline’s <code class="docutils literal notranslate"><span class="pre">named_steps</span></code> attribute allows accessing steps by name with tab
completion in interactive environments:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">pipe</span><span class="o">.</span><span class="n">named_steps</span><span class="o">.</span><span class="n">reduce_dim</span> <span class="ow">is</span> <span class="n">pipe</span><span class="p">[</span><span class="s1">'reduce_dim'</span><span class="p">]</span>
<span class="go">True</span>
</pre></div>
</div>
<p>A sub-pipeline can also be extracted using the slicing notation commonly used
for Python Sequences such as lists or strings (although only a step of 1 is
permitted). This is convenient for performing only some of the transformations
(or their inverse):</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">pipe</span><span class="p">[:</span><span class="mi">1</span><span class="p">]</span>
<span class="go">Pipeline(steps=[('reduce_dim', PCA())])</span>
<span class="gp">>>> </span><span class="n">pipe</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">:]</span>
<span class="go">Pipeline(steps=[('clf', SVC())])</span>
</pre></div>
</div>
</section>
<section id="nested-parameters">
<span id="pipeline-nested-parameters"></span><h4><span class="section-number">6.1.1.1.3. </span>Nested parameters<a class="headerlink" href="#nested-parameters" title="Permalink to this heading">¶</a></h4>
<p>Parameters of the estimators in the pipeline can be accessed using the
<code class="docutils literal notranslate"><span class="pre"><estimator>__<parameter></span></code> syntax:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">pipe</span><span class="o">.</span><span class="n">set_params</span><span class="p">(</span><span class="n">clf__C</span><span class="o">=</span><span class="mi">10</span><span class="p">)</span>
<span class="go">Pipeline(steps=[('reduce_dim', PCA()), ('clf', SVC(C=10))])</span>
</pre></div>
</div>
<p>This is particularly important for doing grid searches:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.model_selection</span> <span class="kn">import</span> <span class="n">GridSearchCV</span>
<span class="gp">>>> </span><span class="n">param_grid</span> <span class="o">=</span> <span class="nb">dict</span><span class="p">(</span><span class="n">reduce_dim__n_components</span><span class="o">=</span><span class="p">[</span><span class="mi">2</span><span class="p">,</span> <span class="mi">5</span><span class="p">,</span> <span class="mi">10</span><span class="p">],</span>
<span class="gp">... </span> <span class="n">clf__C</span><span class="o">=</span><span class="p">[</span><span class="mf">0.1</span><span class="p">,</span> <span class="mi">10</span><span class="p">,</span> <span class="mi">100</span><span class="p">])</span>
<span class="gp">>>> </span><span class="n">grid_search</span> <span class="o">=</span> <span class="n">GridSearchCV</span><span class="p">(</span><span class="n">pipe</span><span class="p">,</span> <span class="n">param_grid</span><span class="o">=</span><span class="n">param_grid</span><span class="p">)</span>
</pre></div>
</div>
<p>Individual steps may also be replaced as parameters, and non-final steps may be
ignored by setting them to <code class="docutils literal notranslate"><span class="pre">'passthrough'</span></code>:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.linear_model</span> <span class="kn">import</span> <span class="n">LogisticRegression</span>
<span class="gp">>>> </span><span class="n">param_grid</span> <span class="o">=</span> <span class="nb">dict</span><span class="p">(</span><span class="n">reduce_dim</span><span class="o">=</span><span class="p">[</span><span class="s1">'passthrough'</span><span class="p">,</span> <span class="n">PCA</span><span class="p">(</span><span class="mi">5</span><span class="p">),</span> <span class="n">PCA</span><span class="p">(</span><span class="mi">10</span><span class="p">)],</span>
<span class="gp">... </span> <span class="n">clf</span><span class="o">=</span><span class="p">[</span><span class="n">SVC</span><span class="p">(),</span> <span class="n">LogisticRegression</span><span class="p">()],</span>
<span class="gp">... </span> <span class="n">clf__C</span><span class="o">=</span><span class="p">[</span><span class="mf">0.1</span><span class="p">,</span> <span class="mi">10</span><span class="p">,</span> <span class="mi">100</span><span class="p">])</span>
<span class="gp">>>> </span><span class="n">grid_search</span> <span class="o">=</span> <span class="n">GridSearchCV</span><span class="p">(</span><span class="n">pipe</span><span class="p">,</span> <span class="n">param_grid</span><span class="o">=</span><span class="n">param_grid</span><span class="p">)</span>
</pre></div>
</div>
<p>The estimators of the pipeline can be retrieved by index:</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">pipe</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
<span class="go">PCA()</span>
</pre></div>
</div>
<p>or by name:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">pipe</span><span class="p">[</span><span class="s1">'reduce_dim'</span><span class="p">]</span>
<span class="go">PCA()</span>
</pre></div>
</div>
<p>To enable model inspection, <a class="reference internal" href="generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">Pipeline</span></code></a> has a
<code class="docutils literal notranslate"><span class="pre">get_feature_names_out()</span></code> method, just like all transformers. You can use
pipeline slicing to get the feature names going into each step:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.datasets</span> <span class="kn">import</span> <span class="n">load_iris</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.feature_selection</span> <span class="kn">import</span> <span class="n">SelectKBest</span>
<span class="gp">>>> </span><span class="n">iris</span> <span class="o">=</span> <span class="n">load_iris</span><span class="p">()</span>
<span class="gp">>>> </span><span class="n">pipe</span> <span class="o">=</span> <span class="n">Pipeline</span><span class="p">(</span><span class="n">steps</span><span class="o">=</span><span class="p">[</span>
<span class="gp">... </span> <span class="p">(</span><span class="s1">'select'</span><span class="p">,</span> <span class="n">SelectKBest</span><span class="p">(</span><span class="n">k</span><span class="o">=</span><span class="mi">2</span><span class="p">)),</span>
<span class="gp">... </span> <span class="p">(</span><span class="s1">'clf'</span><span class="p">,</span> <span class="n">LogisticRegression</span><span class="p">())])</span>
<span class="gp">>>> </span><span class="n">pipe</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">iris</span><span class="o">.</span><span class="n">data</span><span class="p">,</span> <span class="n">iris</span><span class="o">.</span><span class="n">target</span><span class="p">)</span>
<span class="go">Pipeline(steps=[('select', SelectKBest(...)), ('clf', LogisticRegression(...))])</span>
<span class="gp">>>> </span><span class="n">pipe</span><span class="p">[:</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span><span class="o">.</span><span class="n">get_feature_names_out</span><span class="p">()</span>
<span class="go">array(['x2', 'x3'], ...)</span>
</pre></div>
</div>
<p>You can also provide custom feature names for the input data using
<code class="docutils literal notranslate"><span class="pre">get_feature_names_out</span></code>:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">pipe</span><span class="p">[:</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span><span class="o">.</span><span class="n">get_feature_names_out</span><span class="p">(</span><span class="n">iris</span><span class="o">.</span><span class="n">feature_names</span><span class="p">)</span>
<span class="go">array(['petal length (cm)', 'petal width (cm)'], ...)</span>
</pre></div>
</div>
<aside class="topic">
<p class="topic-title">Examples:</p>
<ul class="simple">
<li><p><a class="reference internal" href="../auto_examples/feature_selection/plot_feature_selection_pipeline.html#sphx-glr-auto-examples-feature-selection-plot-feature-selection-pipeline-py"><span class="std std-ref">Pipeline ANOVA SVM</span></a></p></li>
<li><p><a class="reference internal" href="../auto_examples/model_selection/plot_grid_search_text_feature_extraction.html#sphx-glr-auto-examples-model-selection-plot-grid-search-text-feature-extraction-py"><span class="std std-ref">Sample pipeline for text feature extraction and evaluation</span></a></p></li>
<li><p><a class="reference internal" href="../auto_examples/compose/plot_digits_pipe.html#sphx-glr-auto-examples-compose-plot-digits-pipe-py"><span class="std std-ref">Pipelining: chaining a PCA and a logistic regression</span></a></p></li>
<li><p><a class="reference internal" href="../auto_examples/miscellaneous/plot_kernel_approximation.html#sphx-glr-auto-examples-miscellaneous-plot-kernel-approximation-py"><span class="std std-ref">Explicit feature map approximation for RBF kernels</span></a></p></li>
<li><p><a class="reference internal" href="../auto_examples/svm/plot_svm_anova.html#sphx-glr-auto-examples-svm-plot-svm-anova-py"><span class="std std-ref">SVM-Anova: SVM with univariate feature selection</span></a></p></li>
<li><p><a class="reference internal" href="../auto_examples/compose/plot_compare_reduction.html#sphx-glr-auto-examples-compose-plot-compare-reduction-py"><span class="std std-ref">Selecting dimensionality reduction with Pipeline and GridSearchCV</span></a></p></li>
<li><p><a class="reference internal" href="../auto_examples/miscellaneous/plot_pipeline_display.html#sphx-glr-auto-examples-miscellaneous-plot-pipeline-display-py"><span class="std std-ref">Displaying Pipelines</span></a></p></li>
</ul>
</aside>
<aside class="topic">
<p class="topic-title">See Also:</p>
<ul class="simple">
<li><p><a class="reference internal" href="grid_search.html#composite-grid-search"><span class="std std-ref">Composite estimators and parameter spaces</span></a></p></li>
</ul>
</aside>
</section>
</section>
<section id="notes">
<h3><span class="section-number">6.1.1.2. </span>Notes<a class="headerlink" href="#notes" title="Permalink to this heading">¶</a></h3>
<p>Calling <code class="docutils literal notranslate"><span class="pre">fit</span></code> on the pipeline is the same as calling <code class="docutils literal notranslate"><span class="pre">fit</span></code> on
each estimator in turn, <code class="docutils literal notranslate"><span class="pre">transform</span></code> the input and pass it on to the next step.
The pipeline has all the methods that the last estimator in the pipeline has,
i.e. if the last estimator is a classifier, the <a class="reference internal" href="generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">Pipeline</span></code></a> can be used
as a classifier. If the last estimator is a transformer, again, so is the
pipeline.</p>
</section>
<section id="caching-transformers-avoid-repeated-computation">
<span id="pipeline-cache"></span><h3><span class="section-number">6.1.1.3. </span>Caching transformers: avoid repeated computation<a class="headerlink" href="#caching-transformers-avoid-repeated-computation" title="Permalink to this heading">¶</a></h3>
<p>Fitting transformers may be computationally expensive. With its
<code class="docutils literal notranslate"><span class="pre">memory</span></code> parameter set, <a class="reference internal" href="generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">Pipeline</span></code></a> will cache each transformer
after calling <code class="docutils literal notranslate"><span class="pre">fit</span></code>.
This feature is used to avoid computing the fit transformers within a pipeline
if the parameters and input data are identical. A typical example is the case of
a grid search in which the transformers can be fitted only once and reused for
each configuration. The last step will never be cached, even if it is a transformer.</p>
<p>The parameter <code class="docutils literal notranslate"><span class="pre">memory</span></code> is needed in order to cache the transformers.
<code class="docutils literal notranslate"><span class="pre">memory</span></code> can be either a string containing the directory where to cache the
transformers or a <a class="reference external" href="https://joblib.readthedocs.io/en/latest/memory.html">joblib.Memory</a>
object:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">tempfile</span> <span class="kn">import</span> <span class="n">mkdtemp</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">shutil</span> <span class="kn">import</span> <span class="n">rmtree</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.decomposition</span> <span class="kn">import</span> <span class="n">PCA</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.svm</span> <span class="kn">import</span> <span class="n">SVC</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.pipeline</span> <span class="kn">import</span> <span class="n">Pipeline</span>
<span class="gp">>>> </span><span class="n">estimators</span> <span class="o">=</span> <span class="p">[(</span><span class="s1">'reduce_dim'</span><span class="p">,</span> <span class="n">PCA</span><span class="p">()),</span> <span class="p">(</span><span class="s1">'clf'</span><span class="p">,</span> <span class="n">SVC</span><span class="p">())]</span>
<span class="gp">>>> </span><span class="n">cachedir</span> <span class="o">=</span> <span class="n">mkdtemp</span><span class="p">()</span>
<span class="gp">>>> </span><span class="n">pipe</span> <span class="o">=</span> <span class="n">Pipeline</span><span class="p">(</span><span class="n">estimators</span><span class="p">,</span> <span class="n">memory</span><span class="o">=</span><span class="n">cachedir</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">pipe</span>
<span class="go">Pipeline(memory=...,</span>
<span class="go"> steps=[('reduce_dim', PCA()), ('clf', SVC())])</span>
<span class="gp">>>> </span><span class="c1"># Clear the cache directory when you don't need it anymore</span>
<span class="gp">>>> </span><span class="n">rmtree</span><span class="p">(</span><span class="n">cachedir</span><span class="p">)</span>
</pre></div>
</div>
<div class="admonition warning">
<p class="admonition-title">Warning</p>
<p><strong>Side effect of caching transformers</strong></p>
<p>Using a <a class="reference internal" href="generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">Pipeline</span></code></a> without cache enabled, it is possible to
inspect the original instance such as:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.datasets</span> <span class="kn">import</span> <span class="n">load_digits</span>
<span class="gp">>>> </span><span class="n">X_digits</span><span class="p">,</span> <span class="n">y_digits</span> <span class="o">=</span> <span class="n">load_digits</span><span class="p">(</span><span class="n">return_X_y</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">pca1</span> <span class="o">=</span> <span class="n">PCA</span><span class="p">()</span>
<span class="gp">>>> </span><span class="n">svm1</span> <span class="o">=</span> <span class="n">SVC</span><span class="p">()</span>
<span class="gp">>>> </span><span class="n">pipe</span> <span class="o">=</span> <span class="n">Pipeline</span><span class="p">([(</span><span class="s1">'reduce_dim'</span><span class="p">,</span> <span class="n">pca1</span><span class="p">),</span> <span class="p">(</span><span class="s1">'clf'</span><span class="p">,</span> <span class="n">svm1</span><span class="p">)])</span>
<span class="gp">>>> </span><span class="n">pipe</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X_digits</span><span class="p">,</span> <span class="n">y_digits</span><span class="p">)</span>
<span class="go">Pipeline(steps=[('reduce_dim', PCA()), ('clf', SVC())])</span>
<span class="gp">>>> </span><span class="c1"># The pca instance can be inspected directly</span>
<span class="gp">>>> </span><span class="nb">print</span><span class="p">(</span><span class="n">pca1</span><span class="o">.</span><span class="n">components_</span><span class="p">)</span>
<span class="go"> [[-1.77484909e-19 ... 4.07058917e-18]]</span>
</pre></div>
</div>
<p>Enabling caching triggers a clone of the transformers before fitting.
Therefore, the transformer instance given to the pipeline cannot be
inspected directly.
In following example, accessing the <a class="reference internal" href="generated/sklearn.decomposition.PCA.html#sklearn.decomposition.PCA" title="sklearn.decomposition.PCA"><code class="xref py py-class docutils literal notranslate"><span class="pre">PCA</span></code></a>
instance <code class="docutils literal notranslate"><span class="pre">pca2</span></code> will raise an <code class="docutils literal notranslate"><span class="pre">AttributeError</span></code> since <code class="docutils literal notranslate"><span class="pre">pca2</span></code> will be an
unfitted transformer.
Instead, use the attribute <code class="docutils literal notranslate"><span class="pre">named_steps</span></code> to inspect estimators within
the pipeline:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">cachedir</span> <span class="o">=</span> <span class="n">mkdtemp</span><span class="p">()</span>
<span class="gp">>>> </span><span class="n">pca2</span> <span class="o">=</span> <span class="n">PCA</span><span class="p">()</span>
<span class="gp">>>> </span><span class="n">svm2</span> <span class="o">=</span> <span class="n">SVC</span><span class="p">()</span>
<span class="gp">>>> </span><span class="n">cached_pipe</span> <span class="o">=</span> <span class="n">Pipeline</span><span class="p">([(</span><span class="s1">'reduce_dim'</span><span class="p">,</span> <span class="n">pca2</span><span class="p">),</span> <span class="p">(</span><span class="s1">'clf'</span><span class="p">,</span> <span class="n">svm2</span><span class="p">)],</span>
<span class="gp">... </span> <span class="n">memory</span><span class="o">=</span><span class="n">cachedir</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">cached_pipe</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X_digits</span><span class="p">,</span> <span class="n">y_digits</span><span class="p">)</span>
<span class="go">Pipeline(memory=...,</span>
<span class="go"> steps=[('reduce_dim', PCA()), ('clf', SVC())])</span>
<span class="gp">>>> </span><span class="nb">print</span><span class="p">(</span><span class="n">cached_pipe</span><span class="o">.</span><span class="n">named_steps</span><span class="p">[</span><span class="s1">'reduce_dim'</span><span class="p">]</span><span class="o">.</span><span class="n">components_</span><span class="p">)</span>
<span class="go"> [[-1.77484909e-19 ... 4.07058917e-18]]</span>
<span class="gp">>>> </span><span class="c1"># Remove the cache directory</span>
<span class="gp">>>> </span><span class="n">rmtree</span><span class="p">(</span><span class="n">cachedir</span><span class="p">)</span>
</pre></div>
</div>
</div>
<aside class="topic">
<p class="topic-title">Examples:</p>
<ul class="simple">
<li><p><a class="reference internal" href="../auto_examples/compose/plot_compare_reduction.html#sphx-glr-auto-examples-compose-plot-compare-reduction-py"><span class="std std-ref">Selecting dimensionality reduction with Pipeline and GridSearchCV</span></a></p></li>
</ul>
</aside>
</section>
</section>
<section id="transforming-target-in-regression">
<span id="transformed-target-regressor"></span><h2><span class="section-number">6.1.2. </span>Transforming target in regression<a class="headerlink" href="#transforming-target-in-regression" title="Permalink to this heading">¶</a></h2>
<p><a class="reference internal" href="generated/sklearn.compose.TransformedTargetRegressor.html#sklearn.compose.TransformedTargetRegressor" title="sklearn.compose.TransformedTargetRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">TransformedTargetRegressor</span></code></a> transforms the
targets <code class="docutils literal notranslate"><span class="pre">y</span></code> before fitting a regression model. The predictions are mapped
back to the original space via an inverse transform. It takes as an argument
the regressor that will be used for prediction, and the transformer that will
be applied to the target variable:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.datasets</span> <span class="kn">import</span> <span class="n">fetch_california_housing</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.compose</span> <span class="kn">import</span> <span class="n">TransformedTargetRegressor</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.preprocessing</span> <span class="kn">import</span> <span class="n">QuantileTransformer</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.linear_model</span> <span class="kn">import</span> <span class="n">LinearRegression</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.model_selection</span> <span class="kn">import</span> <span class="n">train_test_split</span>
<span class="gp">>>> </span><span class="n">X</span><span class="p">,</span> <span class="n">y</span> <span class="o">=</span> <span class="n">fetch_california_housing</span><span class="p">(</span><span class="n">return_X_y</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">X</span><span class="p">,</span> <span class="n">y</span> <span class="o">=</span> <span class="n">X</span><span class="p">[:</span><span class="mi">2000</span><span class="p">,</span> <span class="p">:],</span> <span class="n">y</span><span class="p">[:</span><span class="mi">2000</span><span class="p">]</span> <span class="c1"># select a subset of data</span>
<span class="gp">>>> </span><span class="n">transformer</span> <span class="o">=</span> <span class="n">QuantileTransformer</span><span class="p">(</span><span class="n">output_distribution</span><span class="o">=</span><span class="s1">'normal'</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">regressor</span> <span class="o">=</span> <span class="n">LinearRegression</span><span class="p">()</span>
<span class="gp">>>> </span><span class="n">regr</span> <span class="o">=</span> <span class="n">TransformedTargetRegressor</span><span class="p">(</span><span class="n">regressor</span><span class="o">=</span><span class="n">regressor</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">transformer</span><span class="o">=</span><span class="n">transformer</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">X_train</span><span class="p">,</span> <span class="n">X_test</span><span class="p">,</span> <span class="n">y_train</span><span class="p">,</span> <span class="n">y_test</span> <span class="o">=</span> <span class="n">train_test_split</span><span class="p">(</span><span class="n">X</span><span class="p">,</span> <span class="n">y</span><span class="p">,</span> <span class="n">random_state</span><span class="o">=</span><span class="mi">0</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">regr</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X_train</span><span class="p">,</span> <span class="n">y_train</span><span class="p">)</span>
<span class="go">TransformedTargetRegressor(...)</span>
<span class="gp">>>> </span><span class="nb">print</span><span class="p">(</span><span class="s1">'R2 score: </span><span class="si">{0:.2f}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">regr</span><span class="o">.</span><span class="n">score</span><span class="p">(</span><span class="n">X_test</span><span class="p">,</span> <span class="n">y_test</span><span class="p">)))</span>
<span class="go">R2 score: 0.61</span>
<span class="gp">>>> </span><span class="n">raw_target_regr</span> <span class="o">=</span> <span class="n">LinearRegression</span><span class="p">()</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X_train</span><span class="p">,</span> <span class="n">y_train</span><span class="p">)</span>
<span class="gp">>>> </span><span class="nb">print</span><span class="p">(</span><span class="s1">'R2 score: </span><span class="si">{0:.2f}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">raw_target_regr</span><span class="o">.</span><span class="n">score</span><span class="p">(</span><span class="n">X_test</span><span class="p">,</span> <span class="n">y_test</span><span class="p">)))</span>
<span class="go">R2 score: 0.59</span>
</pre></div>
</div>
<p>For simple transformations, instead of a Transformer object, a pair of
functions can be passed, defining the transformation and its inverse mapping:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="k">def</span> <span class="nf">func</span><span class="p">(</span><span class="n">x</span><span class="p">):</span>
<span class="gp">... </span> <span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">log</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
<span class="gp">>>> </span><span class="k">def</span> <span class="nf">inverse_func</span><span class="p">(</span><span class="n">x</span><span class="p">):</span>
<span class="gp">... </span> <span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">exp</span><span class="p">(</span><span class="n">x</span><span class="p">)</span>
</pre></div>
</div>
<p>Subsequently, the object is created as:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">regr</span> <span class="o">=</span> <span class="n">TransformedTargetRegressor</span><span class="p">(</span><span class="n">regressor</span><span class="o">=</span><span class="n">regressor</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">func</span><span class="o">=</span><span class="n">func</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">inverse_func</span><span class="o">=</span><span class="n">inverse_func</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">regr</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X_train</span><span class="p">,</span> <span class="n">y_train</span><span class="p">)</span>
<span class="go">TransformedTargetRegressor(...)</span>
<span class="gp">>>> </span><span class="nb">print</span><span class="p">(</span><span class="s1">'R2 score: </span><span class="si">{0:.2f}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">regr</span><span class="o">.</span><span class="n">score</span><span class="p">(</span><span class="n">X_test</span><span class="p">,</span> <span class="n">y_test</span><span class="p">)))</span>
<span class="go">R2 score: 0.51</span>
</pre></div>
</div>
<p>By default, the provided functions are checked at each fit to be the inverse of
each other. However, it is possible to bypass this checking by setting
<code class="docutils literal notranslate"><span class="pre">check_inverse</span></code> to <code class="docutils literal notranslate"><span class="pre">False</span></code>:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="k">def</span> <span class="nf">inverse_func</span><span class="p">(</span><span class="n">x</span><span class="p">):</span>
<span class="gp">... </span> <span class="k">return</span> <span class="n">x</span>
<span class="gp">>>> </span><span class="n">regr</span> <span class="o">=</span> <span class="n">TransformedTargetRegressor</span><span class="p">(</span><span class="n">regressor</span><span class="o">=</span><span class="n">regressor</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">func</span><span class="o">=</span><span class="n">func</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">inverse_func</span><span class="o">=</span><span class="n">inverse_func</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">check_inverse</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">regr</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X_train</span><span class="p">,</span> <span class="n">y_train</span><span class="p">)</span>
<span class="go">TransformedTargetRegressor(...)</span>
<span class="gp">>>> </span><span class="nb">print</span><span class="p">(</span><span class="s1">'R2 score: </span><span class="si">{0:.2f}</span><span class="s1">'</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">regr</span><span class="o">.</span><span class="n">score</span><span class="p">(</span><span class="n">X_test</span><span class="p">,</span> <span class="n">y_test</span><span class="p">)))</span>
<span class="go">R2 score: -1.57</span>
</pre></div>
</div>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>The transformation can be triggered by setting either <code class="docutils literal notranslate"><span class="pre">transformer</span></code> or the
pair of functions <code class="docutils literal notranslate"><span class="pre">func</span></code> and <code class="docutils literal notranslate"><span class="pre">inverse_func</span></code>. However, setting both
options will raise an error.</p>
</div>
<aside class="topic">
<p class="topic-title">Examples:</p>
<ul class="simple">
<li><p><a class="reference internal" href="../auto_examples/compose/plot_transformed_target.html#sphx-glr-auto-examples-compose-plot-transformed-target-py"><span class="std std-ref">Effect of transforming the targets in regression model</span></a></p></li>
</ul>
</aside>
</section>
<section id="featureunion-composite-feature-spaces">
<span id="feature-union"></span><h2><span class="section-number">6.1.3. </span>FeatureUnion: composite feature spaces<a class="headerlink" href="#featureunion-composite-feature-spaces" title="Permalink to this heading">¶</a></h2>
<p><a class="reference internal" href="generated/sklearn.pipeline.FeatureUnion.html#sklearn.pipeline.FeatureUnion" title="sklearn.pipeline.FeatureUnion"><code class="xref py py-class docutils literal notranslate"><span class="pre">FeatureUnion</span></code></a> combines several transformer objects into a new
transformer that combines their output. A <a class="reference internal" href="generated/sklearn.pipeline.FeatureUnion.html#sklearn.pipeline.FeatureUnion" title="sklearn.pipeline.FeatureUnion"><code class="xref py py-class docutils literal notranslate"><span class="pre">FeatureUnion</span></code></a> takes
a list of transformer objects. During fitting, each of these
is fit to the data independently. The transformers are applied in parallel,
and the feature matrices they output are concatenated side-by-side into a
larger matrix.</p>
<p>When you want to apply different transformations to each field of the data,
see the related class <a class="reference internal" href="generated/sklearn.compose.ColumnTransformer.html#sklearn.compose.ColumnTransformer" title="sklearn.compose.ColumnTransformer"><code class="xref py py-class docutils literal notranslate"><span class="pre">ColumnTransformer</span></code></a>
(see <a class="reference internal" href="#column-transformer"><span class="std std-ref">user guide</span></a>).</p>
<p><a class="reference internal" href="generated/sklearn.pipeline.FeatureUnion.html#sklearn.pipeline.FeatureUnion" title="sklearn.pipeline.FeatureUnion"><code class="xref py py-class docutils literal notranslate"><span class="pre">FeatureUnion</span></code></a> serves the same purposes as <a class="reference internal" href="generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">Pipeline</span></code></a> -
convenience and joint parameter estimation and validation.</p>
<p><a class="reference internal" href="generated/sklearn.pipeline.FeatureUnion.html#sklearn.pipeline.FeatureUnion" title="sklearn.pipeline.FeatureUnion"><code class="xref py py-class docutils literal notranslate"><span class="pre">FeatureUnion</span></code></a> and <a class="reference internal" href="generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">Pipeline</span></code></a> can be combined to
create complex models.</p>
<p>(A <a class="reference internal" href="generated/sklearn.pipeline.FeatureUnion.html#sklearn.pipeline.FeatureUnion" title="sklearn.pipeline.FeatureUnion"><code class="xref py py-class docutils literal notranslate"><span class="pre">FeatureUnion</span></code></a> has no way of checking whether two transformers
might produce identical features. It only produces a union when the
feature sets are disjoint, and making sure they are is the caller’s
responsibility.)</p>
<section id="id1">
<h3><span class="section-number">6.1.3.1. </span>Usage<a class="headerlink" href="#id1" title="Permalink to this heading">¶</a></h3>
<p>A <a class="reference internal" href="generated/sklearn.pipeline.FeatureUnion.html#sklearn.pipeline.FeatureUnion" title="sklearn.pipeline.FeatureUnion"><code class="xref py py-class docutils literal notranslate"><span class="pre">FeatureUnion</span></code></a> is built using a list of <code class="docutils literal notranslate"><span class="pre">(key,</span> <span class="pre">value)</span></code> pairs,
where the <code class="docutils literal notranslate"><span class="pre">key</span></code> is the name you want to give to a given transformation
(an arbitrary string; it only serves as an identifier)
and <code class="docutils literal notranslate"><span class="pre">value</span></code> is an estimator object:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.pipeline</span> <span class="kn">import</span> <span class="n">FeatureUnion</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.decomposition</span> <span class="kn">import</span> <span class="n">PCA</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.decomposition</span> <span class="kn">import</span> <span class="n">KernelPCA</span>
<span class="gp">>>> </span><span class="n">estimators</span> <span class="o">=</span> <span class="p">[(</span><span class="s1">'linear_pca'</span><span class="p">,</span> <span class="n">PCA</span><span class="p">()),</span> <span class="p">(</span><span class="s1">'kernel_pca'</span><span class="p">,</span> <span class="n">KernelPCA</span><span class="p">())]</span>
<span class="gp">>>> </span><span class="n">combined</span> <span class="o">=</span> <span class="n">FeatureUnion</span><span class="p">(</span><span class="n">estimators</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">combined</span>
<span class="go">FeatureUnion(transformer_list=[('linear_pca', PCA()),</span>
<span class="go"> ('kernel_pca', KernelPCA())])</span>
</pre></div>
</div>
<p>Like pipelines, feature unions have a shorthand constructor called
<a class="reference internal" href="generated/sklearn.pipeline.make_union.html#sklearn.pipeline.make_union" title="sklearn.pipeline.make_union"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_union</span></code></a> that does not require explicit naming of the components.</p>
<p>Like <code class="docutils literal notranslate"><span class="pre">Pipeline</span></code>, individual steps may be replaced using <code class="docutils literal notranslate"><span class="pre">set_params</span></code>,
and ignored by setting to <code class="docutils literal notranslate"><span class="pre">'drop'</span></code>:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">combined</span><span class="o">.</span><span class="n">set_params</span><span class="p">(</span><span class="n">kernel_pca</span><span class="o">=</span><span class="s1">'drop'</span><span class="p">)</span>
<span class="go">FeatureUnion(transformer_list=[('linear_pca', PCA()),</span>
<span class="go"> ('kernel_pca', 'drop')])</span>
</pre></div>
</div>
<aside class="topic">
<p class="topic-title">Examples:</p>
<ul class="simple">
<li><p><a class="reference internal" href="../auto_examples/compose/plot_feature_union.html#sphx-glr-auto-examples-compose-plot-feature-union-py"><span class="std std-ref">Concatenating multiple feature extraction methods</span></a></p></li>
</ul>
</aside>
</section>
</section>
<section id="columntransformer-for-heterogeneous-data">
<span id="column-transformer"></span><h2><span class="section-number">6.1.4. </span>ColumnTransformer for heterogeneous data<a class="headerlink" href="#columntransformer-for-heterogeneous-data" title="Permalink to this heading">¶</a></h2>
<p>Many datasets contain features of different types, say text, floats, and dates,
where each type of feature requires separate preprocessing or feature
extraction steps. Often it is easiest to preprocess data before applying
scikit-learn methods, for example using <a class="reference external" href="https://pandas.pydata.org/">pandas</a>.
Processing your data before passing it to scikit-learn might be problematic for
one of the following reasons:</p>
<ol class="arabic simple">
<li><p>Incorporating statistics from test data into the preprocessors makes
cross-validation scores unreliable (known as <em>data leakage</em>),
for example in the case of scalers or imputing missing values.</p></li>
<li><p>You may want to include the parameters of the preprocessors in a
<a class="reference internal" href="grid_search.html#grid-search"><span class="std std-ref">parameter search</span></a>.</p></li>
</ol>
<p>The <a class="reference internal" href="generated/sklearn.compose.ColumnTransformer.html#sklearn.compose.ColumnTransformer" title="sklearn.compose.ColumnTransformer"><code class="xref py py-class docutils literal notranslate"><span class="pre">ColumnTransformer</span></code></a> helps performing different
transformations for different columns of the data, within a
<a class="reference internal" href="generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline" title="sklearn.pipeline.Pipeline"><code class="xref py py-class docutils literal notranslate"><span class="pre">Pipeline</span></code></a> that is safe from data leakage and that can
be parametrized. <a class="reference internal" href="generated/sklearn.compose.ColumnTransformer.html#sklearn.compose.ColumnTransformer" title="sklearn.compose.ColumnTransformer"><code class="xref py py-class docutils literal notranslate"><span class="pre">ColumnTransformer</span></code></a> works on
arrays, sparse matrices, and
<a class="reference external" href="https://pandas.pydata.org/pandas-docs/stable/">pandas DataFrames</a>.</p>
<p>To each column, a different transformation can be applied, such as
preprocessing or a specific feature extraction method:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">import</span> <span class="nn">pandas</span> <span class="k">as</span> <span class="nn">pd</span>
<span class="gp">>>> </span><span class="n">X</span> <span class="o">=</span> <span class="n">pd</span><span class="o">.</span><span class="n">DataFrame</span><span class="p">(</span>
<span class="gp">... </span> <span class="p">{</span><span class="s1">'city'</span><span class="p">:</span> <span class="p">[</span><span class="s1">'London'</span><span class="p">,</span> <span class="s1">'London'</span><span class="p">,</span> <span class="s1">'Paris'</span><span class="p">,</span> <span class="s1">'Sallisaw'</span><span class="p">],</span>
<span class="gp">... </span> <span class="s1">'title'</span><span class="p">:</span> <span class="p">[</span><span class="s2">"His Last Bow"</span><span class="p">,</span> <span class="s2">"How Watson Learned the Trick"</span><span class="p">,</span>
<span class="gp">... </span> <span class="s2">"A Moveable Feast"</span><span class="p">,</span> <span class="s2">"The Grapes of Wrath"</span><span class="p">],</span>
<span class="gp">... </span> <span class="s1">'expert_rating'</span><span class="p">:</span> <span class="p">[</span><span class="mi">5</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">4</span><span class="p">,</span> <span class="mi">5</span><span class="p">],</span>
<span class="gp">... </span> <span class="s1">'user_rating'</span><span class="p">:</span> <span class="p">[</span><span class="mi">4</span><span class="p">,</span> <span class="mi">5</span><span class="p">,</span> <span class="mi">4</span><span class="p">,</span> <span class="mi">3</span><span class="p">]})</span>
</pre></div>
</div>
<p>For this data, we might want to encode the <code class="docutils literal notranslate"><span class="pre">'city'</span></code> column as a categorical
variable using <a class="reference internal" href="generated/sklearn.preprocessing.OneHotEncoder.html#sklearn.preprocessing.OneHotEncoder" title="sklearn.preprocessing.OneHotEncoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">OneHotEncoder</span></code></a> but apply a
<a class="reference internal" href="generated/sklearn.feature_extraction.text.CountVectorizer.html#sklearn.feature_extraction.text.CountVectorizer" title="sklearn.feature_extraction.text.CountVectorizer"><code class="xref py py-class docutils literal notranslate"><span class="pre">CountVectorizer</span></code></a> to the <code class="docutils literal notranslate"><span class="pre">'title'</span></code> column.
As we might use multiple feature extraction methods on the same column, we give
each transformer a unique name, say <code class="docutils literal notranslate"><span class="pre">'city_category'</span></code> and <code class="docutils literal notranslate"><span class="pre">'title_bow'</span></code>.
By default, the remaining rating columns are ignored (<code class="docutils literal notranslate"><span class="pre">remainder='drop'</span></code>):</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.compose</span> <span class="kn">import</span> <span class="n">ColumnTransformer</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.feature_extraction.text</span> <span class="kn">import</span> <span class="n">CountVectorizer</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.preprocessing</span> <span class="kn">import</span> <span class="n">OneHotEncoder</span>
<span class="gp">>>> </span><span class="n">column_trans</span> <span class="o">=</span> <span class="n">ColumnTransformer</span><span class="p">(</span>
<span class="gp">... </span> <span class="p">[(</span><span class="s1">'categories'</span><span class="p">,</span> <span class="n">OneHotEncoder</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="s1">'int'</span><span class="p">),</span> <span class="p">[</span><span class="s1">'city'</span><span class="p">]),</span>
<span class="gp">... </span> <span class="p">(</span><span class="s1">'title_bow'</span><span class="p">,</span> <span class="n">CountVectorizer</span><span class="p">(),</span> <span class="s1">'title'</span><span class="p">)],</span>
<span class="gp">... </span> <span class="n">remainder</span><span class="o">=</span><span class="s1">'drop'</span><span class="p">,</span> <span class="n">verbose_feature_names_out</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">column_trans</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X</span><span class="p">)</span>
<span class="go">ColumnTransformer(transformers=[('categories', OneHotEncoder(dtype='int'),</span>
<span class="go"> ['city']),</span>
<span class="go"> ('title_bow', CountVectorizer(), 'title')],</span>
<span class="go"> verbose_feature_names_out=False)</span>
<span class="gp">>>> </span><span class="n">column_trans</span><span class="o">.</span><span class="n">get_feature_names_out</span><span class="p">()</span>
<span class="go">array(['city_London', 'city_Paris', 'city_Sallisaw', 'bow', 'feast',</span>
<span class="go">'grapes', 'his', 'how', 'last', 'learned', 'moveable', 'of', 'the',</span>
<span class="go"> 'trick', 'watson', 'wrath'], ...)</span>
<span class="gp">>>> </span><span class="n">column_trans</span><span class="o">.</span><span class="n">transform</span><span class="p">(</span><span class="n">X</span><span class="p">)</span><span class="o">.</span><span class="n">toarray</span><span class="p">()</span>
<span class="go">array([[1, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0],</span>
<span class="go"> [1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0],</span>
<span class="go"> [0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0],</span>
<span class="go"> [0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1]]...)</span>
</pre></div>
</div>
<p>In the above example, the
<a class="reference internal" href="generated/sklearn.feature_extraction.text.CountVectorizer.html#sklearn.feature_extraction.text.CountVectorizer" title="sklearn.feature_extraction.text.CountVectorizer"><code class="xref py py-class docutils literal notranslate"><span class="pre">CountVectorizer</span></code></a> expects a 1D array as
input and therefore the columns were specified as a string (<code class="docutils literal notranslate"><span class="pre">'title'</span></code>).
However, <a class="reference internal" href="generated/sklearn.preprocessing.OneHotEncoder.html#sklearn.preprocessing.OneHotEncoder" title="sklearn.preprocessing.OneHotEncoder"><code class="xref py py-class docutils literal notranslate"><span class="pre">OneHotEncoder</span></code></a>
as most of other transformers expects 2D data, therefore in that case you need
to specify the column as a list of strings (<code class="docutils literal notranslate"><span class="pre">['city']</span></code>).</p>
<p>Apart from a scalar or a single item list, the column selection can be specified
as a list of multiple items, an integer array, a slice, a boolean mask, or
with a <a class="reference internal" href="generated/sklearn.compose.make_column_selector.html#sklearn.compose.make_column_selector" title="sklearn.compose.make_column_selector"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_column_selector</span></code></a>. The
<a class="reference internal" href="generated/sklearn.compose.make_column_selector.html#sklearn.compose.make_column_selector" title="sklearn.compose.make_column_selector"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_column_selector</span></code></a> is used to select columns based
on data type or column name:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.preprocessing</span> <span class="kn">import</span> <span class="n">StandardScaler</span>
<span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.compose</span> <span class="kn">import</span> <span class="n">make_column_selector</span>
<span class="gp">>>> </span><span class="n">ct</span> <span class="o">=</span> <span class="n">ColumnTransformer</span><span class="p">([</span>
<span class="gp">... </span> <span class="p">(</span><span class="s1">'scale'</span><span class="p">,</span> <span class="n">StandardScaler</span><span class="p">(),</span>
<span class="gp">... </span> <span class="n">make_column_selector</span><span class="p">(</span><span class="n">dtype_include</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">number</span><span class="p">)),</span>
<span class="gp">... </span> <span class="p">(</span><span class="s1">'onehot'</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">OneHotEncoder</span><span class="p">(),</span>
<span class="gp">... </span> <span class="n">make_column_selector</span><span class="p">(</span><span class="n">pattern</span><span class="o">=</span><span class="s1">'city'</span><span class="p">,</span> <span class="n">dtype_include</span><span class="o">=</span><span class="nb">object</span><span class="p">))])</span>
<span class="gp">>>> </span><span class="n">ct</span><span class="o">.</span><span class="n">fit_transform</span><span class="p">(</span><span class="n">X</span><span class="p">)</span>
<span class="go">array([[ 0.904..., 0. , 1. , 0. , 0. ],</span>
<span class="go"> [-1.507..., 1.414..., 1. , 0. , 0. ],</span>
<span class="go"> [-0.301..., 0. , 0. , 1. , 0. ],</span>
<span class="go"> [ 0.904..., -1.414..., 0. , 0. , 1. ]])</span>
</pre></div>
</div>
<p>Strings can reference columns if the input is a DataFrame, integers are always
interpreted as the positional columns.</p>
<p>We can keep the remaining rating columns by setting
<code class="docutils literal notranslate"><span class="pre">remainder='passthrough'</span></code>. The values are appended to the end of the
transformation:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">column_trans</span> <span class="o">=</span> <span class="n">ColumnTransformer</span><span class="p">(</span>
<span class="gp">... </span> <span class="p">[(</span><span class="s1">'city_category'</span><span class="p">,</span> <span class="n">OneHotEncoder</span><span class="p">(</span><span class="n">dtype</span><span class="o">=</span><span class="s1">'int'</span><span class="p">),[</span><span class="s1">'city'</span><span class="p">]),</span>
<span class="gp">... </span> <span class="p">(</span><span class="s1">'title_bow'</span><span class="p">,</span> <span class="n">CountVectorizer</span><span class="p">(),</span> <span class="s1">'title'</span><span class="p">)],</span>
<span class="gp">... </span> <span class="n">remainder</span><span class="o">=</span><span class="s1">'passthrough'</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">column_trans</span><span class="o">.</span><span class="n">fit_transform</span><span class="p">(</span><span class="n">X</span><span class="p">)</span>
<span class="go">array([[1, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 5, 4],</span>
<span class="go"> [1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 3, 5],</span>
<span class="go"> [0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 4, 4],</span>
<span class="go"> [0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 5, 3]]...)</span>
</pre></div>
</div>
<p>The <code class="docutils literal notranslate"><span class="pre">remainder</span></code> parameter can be set to an estimator to transform the
remaining rating columns. The transformed values are appended to the end of
the transformation:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.preprocessing</span> <span class="kn">import</span> <span class="n">MinMaxScaler</span>
<span class="gp">>>> </span><span class="n">column_trans</span> <span class="o">=</span> <span class="n">ColumnTransformer</span><span class="p">(</span>
<span class="gp">... </span> <span class="p">[(</span><span class="s1">'city_category'</span><span class="p">,</span> <span class="n">OneHotEncoder</span><span class="p">(),</span> <span class="p">[</span><span class="s1">'city'</span><span class="p">]),</span>
<span class="gp">... </span> <span class="p">(</span><span class="s1">'title_bow'</span><span class="p">,</span> <span class="n">CountVectorizer</span><span class="p">(),</span> <span class="s1">'title'</span><span class="p">)],</span>
<span class="gp">... </span> <span class="n">remainder</span><span class="o">=</span><span class="n">MinMaxScaler</span><span class="p">())</span>
<span class="gp">>>> </span><span class="n">column_trans</span><span class="o">.</span><span class="n">fit_transform</span><span class="p">(</span><span class="n">X</span><span class="p">)[:,</span> <span class="o">-</span><span class="mi">2</span><span class="p">:]</span>
<span class="go">array([[1. , 0.5],</span>
<span class="go"> [0. , 1. ],</span>
<span class="go"> [0.5, 0.5],</span>
<span class="go"> [1. , 0. ]])</span>
</pre></div>
</div>
<p id="make-column-transformer">The <a class="reference internal" href="generated/sklearn.compose.make_column_transformer.html#sklearn.compose.make_column_transformer" title="sklearn.compose.make_column_transformer"><code class="xref py py-func docutils literal notranslate"><span class="pre">make_column_transformer</span></code></a> function is available
to more easily create a <a class="reference internal" href="generated/sklearn.compose.ColumnTransformer.html#sklearn.compose.ColumnTransformer" title="sklearn.compose.ColumnTransformer"><code class="xref py py-class docutils literal notranslate"><span class="pre">ColumnTransformer</span></code></a> object.
Specifically, the names will be given automatically. The equivalent for the
above example would be:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.compose</span> <span class="kn">import</span> <span class="n">make_column_transformer</span>
<span class="gp">>>> </span><span class="n">column_trans</span> <span class="o">=</span> <span class="n">make_column_transformer</span><span class="p">(</span>
<span class="gp">... </span> <span class="p">(</span><span class="n">OneHotEncoder</span><span class="p">(),</span> <span class="p">[</span><span class="s1">'city'</span><span class="p">]),</span>
<span class="gp">... </span> <span class="p">(</span><span class="n">CountVectorizer</span><span class="p">(),</span> <span class="s1">'title'</span><span class="p">),</span>
<span class="gp">... </span> <span class="n">remainder</span><span class="o">=</span><span class="n">MinMaxScaler</span><span class="p">())</span>
<span class="gp">>>> </span><span class="n">column_trans</span>
<span class="go">ColumnTransformer(remainder=MinMaxScaler(),</span>
<span class="go"> transformers=[('onehotencoder', OneHotEncoder(), ['city']),</span>
<span class="go"> ('countvectorizer', CountVectorizer(),</span>
<span class="go"> 'title')])</span>
</pre></div>
</div>
<p>If <a class="reference internal" href="generated/sklearn.compose.ColumnTransformer.html#sklearn.compose.ColumnTransformer" title="sklearn.compose.ColumnTransformer"><code class="xref py py-class docutils literal notranslate"><span class="pre">ColumnTransformer</span></code></a> is fitted with a dataframe
and the dataframe only has string column names, then transforming a dataframe
will use the column names to select the columns:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">ct</span> <span class="o">=</span> <span class="n">ColumnTransformer</span><span class="p">(</span>
<span class="gp">... </span> <span class="p">[(</span><span class="s2">"scale"</span><span class="p">,</span> <span class="n">StandardScaler</span><span class="p">(),</span> <span class="p">[</span><span class="s2">"expert_rating"</span><span class="p">])])</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X</span><span class="p">)</span>
<span class="gp">>>> </span><span class="n">X_new</span> <span class="o">=</span> <span class="n">pd</span><span class="o">.</span><span class="n">DataFrame</span><span class="p">({</span><span class="s2">"expert_rating"</span><span class="p">:</span> <span class="p">[</span><span class="mi">5</span><span class="p">,</span> <span class="mi">6</span><span class="p">,</span> <span class="mi">1</span><span class="p">],</span>
<span class="gp">... </span> <span class="s2">"ignored_new_col"</span><span class="p">:</span> <span class="p">[</span><span class="mf">1.2</span><span class="p">,</span> <span class="mf">0.3</span><span class="p">,</span> <span class="o">-</span><span class="mf">0.1</span><span class="p">]})</span>
<span class="gp">>>> </span><span class="n">ct</span><span class="o">.</span><span class="n">transform</span><span class="p">(</span><span class="n">X_new</span><span class="p">)</span>
<span class="go">array([[ 0.9...],</span>
<span class="go"> [ 2.1...],</span>
<span class="go"> [-3.9...]])</span>
</pre></div>
</div>
</section>
<section id="visualizing-composite-estimators">
<span id="id2"></span><h2><span class="section-number">6.1.5. </span>Visualizing Composite Estimators<a class="headerlink" href="#visualizing-composite-estimators" title="Permalink to this heading">¶</a></h2>
<p>Estimators are displayed with an HTML representation when shown in a
jupyter notebook. This is useful to diagnose or visualize a Pipeline with
many estimators. This visualization is activated by default:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="n">column_trans</span>
</pre></div>
</div>
<p>It can be deactivated by setting the <code class="docutils literal notranslate"><span class="pre">display</span></code> option in <a class="reference internal" href="generated/sklearn.set_config.html#sklearn.set_config" title="sklearn.set_config"><code class="xref py py-func docutils literal notranslate"><span class="pre">set_config</span></code></a>
to ‘text’:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn</span> <span class="kn">import</span> <span class="n">set_config</span>
<span class="gp">>>> </span><span class="n">set_config</span><span class="p">(</span><span class="n">display</span><span class="o">=</span><span class="s1">'text'</span><span class="p">)</span>
<span class="gp">>>> </span><span class="c1"># displays text representation in a jupyter context</span>
<span class="gp">>>> </span><span class="n">column_trans</span>
</pre></div>
</div>
<p>An example of the HTML output can be seen in the
<strong>HTML representation of Pipeline</strong> section of
<a class="reference internal" href="../auto_examples/compose/plot_column_transformer_mixed_types.html#sphx-glr-auto-examples-compose-plot-column-transformer-mixed-types-py"><span class="std std-ref">Column Transformer with Mixed Types</span></a>.
As an alternative, the HTML can be written to a file using
<a class="reference internal" href="generated/sklearn.utils.estimator_html_repr.html#sklearn.utils.estimator_html_repr" title="sklearn.utils.estimator_html_repr"><code class="xref py py-func docutils literal notranslate"><span class="pre">estimator_html_repr</span></code></a>:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">>>> </span><span class="kn">from</span> <span class="nn">sklearn.utils</span> <span class="kn">import</span> <span class="n">estimator_html_repr</span>
<span class="gp">>>> </span><span class="k">with</span> <span class="nb">open</span><span class="p">(</span><span class="s1">'my_estimator.html'</span><span class="p">,</span> <span class="s1">'w'</span><span class="p">)</span> <span class="k">as</span> <span class="n">f</span><span class="p">:</span>
<span class="gp">... </span> <span class="n">f</span><span class="o">.</span><span class="n">write</span><span class="p">(</span><span class="n">estimator_html_repr</span><span class="p">(</span><span class="n">clf</span><span class="p">))</span>
</pre></div>
</div>
<aside class="topic">
<p class="topic-title">Examples:</p>
<ul class="simple">
<li><p><a class="reference internal" href="../auto_examples/compose/plot_column_transformer.html#sphx-glr-auto-examples-compose-plot-column-transformer-py"><span class="std std-ref">Column Transformer with Heterogeneous Data Sources</span></a></p></li>
<li><p><a class="reference internal" href="../auto_examples/compose/plot_column_transformer_mixed_types.html#sphx-glr-auto-examples-compose-plot-column-transformer-mixed-types-py"><span class="std std-ref">Column Transformer with Mixed Types</span></a></p></li>
</ul>
</aside>
</section>
</section>
</div>
<div class="container">
<footer class="sk-content-footer">
© 2007 - 2023, scikit-learn developers (BSD License).
<a href="../_sources/modules/compose.rst.txt" rel="nofollow">Show this page source</a>
</footer>
</div>
</div>
</div>
<script src="../_static/js/vendor/bootstrap.min.js"></script>
<script>
window.ga=window.ga||function(){(ga.q=ga.q||[]).push(arguments)};ga.l=+new Date;
ga('create', 'UA-22606712-2', 'auto');
ga('set', 'anonymizeIp', true);
ga('send', 'pageview');
</script>
<script async src='https://www.google-analytics.com/analytics.js'></script>
<script defer data-domain="scikit-learn.org" src="https://views.scientific-python.org/js/script.js">
</script>
<script src="../_static/clipboard.min.js"></script>
<script src="../_static/copybutton.js"></script>
<script>
$(document).ready(function() {
/* Add a [>>>] button on the top-right corner of code samples to hide
* the >>> and ... prompts and the output and thus make the code
* copyable. */
var div = $('.highlight-python .highlight,' +
'.highlight-python3 .highlight,' +
'.highlight-pycon .highlight,' +
'.highlight-default .highlight')
var pre = div.find('pre');
// get the styles from the current theme
pre.parent().parent().css('position', 'relative');
// create and add the button to all the code blocks that contain >>>
div.each(function(index) {
var jthis = $(this);
// tracebacks (.gt) contain bare text elements that need to be
// wrapped in a span to work with .nextUntil() (see later)
jthis.find('pre:has(.gt)').contents().filter(function() {
return ((this.nodeType == 3) && (this.data.trim().length > 0));
}).wrap('<span>');
});
/*** Add permalink buttons next to glossary terms ***/
$('dl.glossary > dt[id]').append(function() {
return ('<a class="headerlink" href="#' +
this.getAttribute('id') +
'" title="Permalink to this term">¶</a>');
});
});
</script>
<script id="MathJax-script" async src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-chtml.js"></script>
<script src="https://scikit-learn.org/versionwarning.js"></script>
</body>
</html>