-
Notifications
You must be signed in to change notification settings - Fork 81
/
Copy pathscaling_strategies.html
590 lines (498 loc) · 32.4 KB
/
scaling_strategies.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
<!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="8.1. Strategies to scale computationally: bigger data" />
<meta property="og:type" content="website" />
<meta property="og:url" content="https://scikit-learn/stable/computing/scaling_strategies.html" />
<meta property="og:site_name" content="scikit-learn" />
<meta property="og:description" content="For some applications the amount of examples, features (or both) and/or the speed at which they need to be processed are challenging for traditional approaches. In these cases scikit-learn has a nu..." />
<meta property="og:image" content="https://scikit-learn/stable/computing/auto_examples/applications/images/sphx_glr_plot_out_of_core_classification_001.png" />
<meta property="og:image:alt" content="accuracy_over_time" />
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>8.1. Strategies to scale computationally: bigger data — scikit-learn 1.1.3 documentation</title>
<link rel="canonical" href="http://scikit-learn.org/stable/computing/scaling_strategies.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/plot_directive.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/jquery.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="../modules/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/v1.1.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="https://scikit-learn.org/dev/developers/index.html" target="_blank" rel="noopener noreferrer">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="../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/v1.1.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="https://scikit-learn.org/dev/developers/index.html" target="_blank" rel="noopener noreferrer">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="../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="sk-sidebar-toc-logo">
<a href="../index.html">
<img
class="sk-brand-img"
src="../_static/scikit-learn-logo-small.png"
alt="logo"/>
</a>
</div>
<div class="btn-group w-100 mb-2" role="group" aria-label="rellinks">
<a href="../computing.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="8. Computing with scikit-learn">Prev</a><a href="../computing.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="8. Computing with scikit-learn">Up</a>
<a href="computational_performance.html" role="button" class="btn sk-btn-rellink py-1" sk-rellink-tooltip="8.2. Computational Performance">Next</a>
</div>
<div class="alert alert-danger p-1 mb-2" role="alert">
<p class="text-center mb-0">
<strong>scikit-learn 1.1.3</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 href="../user_guide.html" class="sk-toc-active">User Guide</a>
</li>
<ul>
<li>
<a href="../supervised_learning.html" class="">1. Supervised learning</a>
</li>
<li>
<a href="../unsupervised_learning.html" class="">2. Unsupervised learning</a>
</li>
<li>
<a href="../model_selection.html" class="">3. Model selection and evaluation</a>
</li>
<li>
<a href="../inspection.html" class="">4. Inspection</a>
</li>
<li>
<a href="../visualizations.html" class="">5. Visualizations</a>
</li>
<li>
<a href="../data_transforms.html" class="">6. Dataset transformations</a>
</li>
<li>
<a href="../datasets.html" class="">7. Dataset loading utilities</a>
</li>
<li>
<a href="../computing.html" class="sk-toc-active">8. Computing with scikit-learn</a>
<ul>
<li class="sk-toctree-l3">
<a href="">8.1. Strategies to scale computationally: bigger data</a>
</li>
<li class="sk-toctree-l3">
<a href="computational_performance.html">8.2. Computational Performance</a>
</li>
<li class="sk-toctree-l3">
<a href="parallelism.html">8.3. Parallelism, resource management, and configuration</a>
</li>
</ul>
</li>
<li>
<a href="../model_persistence.html" class="">9. Model persistence</a>
</li>
<li>
<a href="../common_pitfalls.html" class="">10. Common pitfalls and recommended practices</a>
</li>
</ul>
</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="strategies-to-scale-computationally-bigger-data">
<span id="scaling-strategies"></span><h1><span class="section-number">8.1. </span>Strategies to scale computationally: bigger data<a class="headerlink" href="#strategies-to-scale-computationally-bigger-data" title="Permalink to this heading">¶</a></h1>
<p>For some applications the amount of examples, features (or both) and/or the
speed at which they need to be processed are challenging for traditional
approaches. In these cases scikit-learn has a number of options you can
consider to make your system scale.</p>
<section id="scaling-with-instances-using-out-of-core-learning">
<h2><span class="section-number">8.1.1. </span>Scaling with instances using out-of-core learning<a class="headerlink" href="#scaling-with-instances-using-out-of-core-learning" title="Permalink to this heading">¶</a></h2>
<p>Out-of-core (or “external memory”) learning is a technique used to learn from
data that cannot fit in a computer’s main memory (RAM).</p>
<p>Here is a sketch of a system designed to achieve this goal:</p>
<blockquote>
<div><ol class="arabic simple">
<li><p>a way to stream instances</p></li>
<li><p>a way to extract features from instances</p></li>
<li><p>an incremental algorithm</p></li>
</ol>
</div></blockquote>
<section id="streaming-instances">
<h3><span class="section-number">8.1.1.1. </span>Streaming instances<a class="headerlink" href="#streaming-instances" title="Permalink to this heading">¶</a></h3>
<p>Basically, 1. may be a reader that yields instances from files on a
hard drive, a database, from a network stream etc. However,
details on how to achieve this are beyond the scope of this documentation.</p>
</section>
<section id="extracting-features">
<h3><span class="section-number">8.1.1.2. </span>Extracting features<a class="headerlink" href="#extracting-features" title="Permalink to this heading">¶</a></h3>
<p>2. could be any relevant way to extract features among the
different <a class="reference internal" href="../modules/feature_extraction.html#feature-extraction"><span class="std std-ref">feature extraction</span></a> methods supported by
scikit-learn. However, when working with data that needs vectorization and
where the set of features or values is not known in advance one should take
explicit care. A good example is text classification where unknown terms are
likely to be found during training. It is possible to use a stateful
vectorizer if making multiple passes over the data is reasonable from an
application point of view. Otherwise, one can turn up the difficulty by using
a stateless feature extractor. Currently the preferred way to do this is to
use the so-called <a class="reference internal" href="../modules/feature_extraction.html#feature-hashing"><span class="std std-ref">hashing trick</span></a> as implemented by
<a class="reference internal" href="../modules/generated/sklearn.feature_extraction.FeatureHasher.html#sklearn.feature_extraction.FeatureHasher" title="sklearn.feature_extraction.FeatureHasher"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.feature_extraction.FeatureHasher</span></code></a> for datasets with categorical
variables represented as list of Python dicts or
<a class="reference internal" href="../modules/generated/sklearn.feature_extraction.text.HashingVectorizer.html#sklearn.feature_extraction.text.HashingVectorizer" title="sklearn.feature_extraction.text.HashingVectorizer"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.feature_extraction.text.HashingVectorizer</span></code></a> for text documents.</p>
</section>
<section id="incremental-learning">
<h3><span class="section-number">8.1.1.3. </span>Incremental learning<a class="headerlink" href="#incremental-learning" title="Permalink to this heading">¶</a></h3>
<p>Finally, for 3. we have a number of options inside scikit-learn. Although not
all algorithms can learn incrementally (i.e. without seeing all the instances
at once), all estimators implementing the <code class="docutils literal notranslate"><span class="pre">partial_fit</span></code> API are candidates.
Actually, the ability to learn incrementally from a mini-batch of instances
(sometimes called “online learning”) is key to out-of-core learning as it
guarantees that at any given time there will be only a small amount of
instances in the main memory. Choosing a good size for the mini-batch that
balances relevancy and memory footprint could involve some tuning <a class="footnote-reference brackets" href="#id2" id="id1" role="doc-noteref"><span class="fn-bracket">[</span>1<span class="fn-bracket">]</span></a>.</p>
<p>Here is a list of incremental estimators for different tasks:</p>
<blockquote>
<div><ul class="simple">
<li><dl class="simple">
<dt>Classification</dt><dd><ul>
<li><p><a class="reference internal" href="../modules/generated/sklearn.naive_bayes.MultinomialNB.html#sklearn.naive_bayes.MultinomialNB" title="sklearn.naive_bayes.MultinomialNB"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.naive_bayes.MultinomialNB</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.naive_bayes.BernoulliNB.html#sklearn.naive_bayes.BernoulliNB" title="sklearn.naive_bayes.BernoulliNB"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.naive_bayes.BernoulliNB</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.linear_model.Perceptron.html#sklearn.linear_model.Perceptron" title="sklearn.linear_model.Perceptron"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.linear_model.Perceptron</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.linear_model.SGDClassifier.html#sklearn.linear_model.SGDClassifier" title="sklearn.linear_model.SGDClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.linear_model.SGDClassifier</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.linear_model.PassiveAggressiveClassifier.html#sklearn.linear_model.PassiveAggressiveClassifier" title="sklearn.linear_model.PassiveAggressiveClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.linear_model.PassiveAggressiveClassifier</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.neural_network.MLPClassifier.html#sklearn.neural_network.MLPClassifier" title="sklearn.neural_network.MLPClassifier"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.neural_network.MLPClassifier</span></code></a></p></li>
</ul>
</dd>
</dl>
</li>
<li><dl class="simple">
<dt>Regression</dt><dd><ul>
<li><p><a class="reference internal" href="../modules/generated/sklearn.linear_model.SGDRegressor.html#sklearn.linear_model.SGDRegressor" title="sklearn.linear_model.SGDRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.linear_model.SGDRegressor</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.linear_model.PassiveAggressiveRegressor.html#sklearn.linear_model.PassiveAggressiveRegressor" title="sklearn.linear_model.PassiveAggressiveRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.linear_model.PassiveAggressiveRegressor</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.neural_network.MLPRegressor.html#sklearn.neural_network.MLPRegressor" title="sklearn.neural_network.MLPRegressor"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.neural_network.MLPRegressor</span></code></a></p></li>
</ul>
</dd>
</dl>
</li>
<li><dl class="simple">
<dt>Clustering</dt><dd><ul>
<li><p><a class="reference internal" href="../modules/generated/sklearn.cluster.MiniBatchKMeans.html#sklearn.cluster.MiniBatchKMeans" title="sklearn.cluster.MiniBatchKMeans"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.cluster.MiniBatchKMeans</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.cluster.Birch.html#sklearn.cluster.Birch" title="sklearn.cluster.Birch"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.cluster.Birch</span></code></a></p></li>
</ul>
</dd>
</dl>
</li>
<li><dl class="simple">
<dt>Decomposition / feature Extraction</dt><dd><ul>
<li><p><a class="reference internal" href="../modules/generated/sklearn.decomposition.MiniBatchDictionaryLearning.html#sklearn.decomposition.MiniBatchDictionaryLearning" title="sklearn.decomposition.MiniBatchDictionaryLearning"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.decomposition.MiniBatchDictionaryLearning</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.decomposition.IncrementalPCA.html#sklearn.decomposition.IncrementalPCA" title="sklearn.decomposition.IncrementalPCA"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.decomposition.IncrementalPCA</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.decomposition.LatentDirichletAllocation.html#sklearn.decomposition.LatentDirichletAllocation" title="sklearn.decomposition.LatentDirichletAllocation"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.decomposition.LatentDirichletAllocation</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.decomposition.MiniBatchNMF.html#sklearn.decomposition.MiniBatchNMF" title="sklearn.decomposition.MiniBatchNMF"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.decomposition.MiniBatchNMF</span></code></a></p></li>
</ul>
</dd>
</dl>
</li>
<li><dl class="simple">
<dt>Preprocessing</dt><dd><ul>
<li><p><a class="reference internal" href="../modules/generated/sklearn.preprocessing.StandardScaler.html#sklearn.preprocessing.StandardScaler" title="sklearn.preprocessing.StandardScaler"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.preprocessing.StandardScaler</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.preprocessing.MinMaxScaler.html#sklearn.preprocessing.MinMaxScaler" title="sklearn.preprocessing.MinMaxScaler"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.preprocessing.MinMaxScaler</span></code></a></p></li>
<li><p><a class="reference internal" href="../modules/generated/sklearn.preprocessing.MaxAbsScaler.html#sklearn.preprocessing.MaxAbsScaler" title="sklearn.preprocessing.MaxAbsScaler"><code class="xref py py-class docutils literal notranslate"><span class="pre">sklearn.preprocessing.MaxAbsScaler</span></code></a></p></li>
</ul>
</dd>
</dl>
</li>
</ul>
</div></blockquote>
<p>For classification, a somewhat important thing to note is that although a
stateless feature extraction routine may be able to cope with new/unseen
attributes, the incremental learner itself may be unable to cope with
new/unseen targets classes. In this case you have to pass all the possible
classes to the first <code class="docutils literal notranslate"><span class="pre">partial_fit</span></code> call using the <code class="docutils literal notranslate"><span class="pre">classes=</span></code> parameter.</p>
<p>Another aspect to consider when choosing a proper algorithm is that not all of
them put the same importance on each example over time. Namely, the
<code class="docutils literal notranslate"><span class="pre">Perceptron</span></code> is still sensitive to badly labeled examples even after many
examples whereas the <code class="docutils literal notranslate"><span class="pre">SGD*</span></code> and <code class="docutils literal notranslate"><span class="pre">PassiveAggressive*</span></code> families are more
robust to this kind of artifacts. Conversely, the latter also tend to give less
importance to remarkably different, yet properly labeled examples when they
come late in the stream as their learning rate decreases over time.</p>
</section>
<section id="examples">
<h3><span class="section-number">8.1.1.4. </span>Examples<a class="headerlink" href="#examples" title="Permalink to this heading">¶</a></h3>
<p>Finally, we have a full-fledged example of
<a class="reference internal" href="../auto_examples/applications/plot_out_of_core_classification.html#sphx-glr-auto-examples-applications-plot-out-of-core-classification-py"><span class="std std-ref">Out-of-core classification of text documents</span></a>. It is aimed at
providing a starting point for people wanting to build out-of-core learning
systems and demonstrates most of the notions discussed above.</p>
<p>Furthermore, it also shows the evolution of the performance of different
algorithms with the number of processed examples.</p>
<p class="centered">
<strong><a class="reference external" href="../auto_examples/applications/plot_out_of_core_classification.html"><img alt="accuracy_over_time" src="../_images/sphx_glr_plot_out_of_core_classification_001.png" style="width: 512.0px; height: 384.0px;" /></a></strong></p><p>Now looking at the computation time of the different parts, we see that the
vectorization is much more expensive than learning itself. From the different
algorithms, <code class="docutils literal notranslate"><span class="pre">MultinomialNB</span></code> is the most expensive, but its overhead can be
mitigated by increasing the size of the mini-batches (exercise: change
<code class="docutils literal notranslate"><span class="pre">minibatch_size</span></code> to 100 and 10000 in the program and compare).</p>
<p class="centered">
<strong><a class="reference external" href="../auto_examples/applications/plot_out_of_core_classification.html"><img alt="computation_time" src="../_images/sphx_glr_plot_out_of_core_classification_003.png" style="width: 512.0px; height: 384.0px;" /></a></strong></p></section>
<section id="notes">
<h3><span class="section-number">8.1.1.5. </span>Notes<a class="headerlink" href="#notes" title="Permalink to this heading">¶</a></h3>
<aside class="footnote-list brackets">
<aside class="footnote brackets" id="id2" role="note">
<span class="label"><span class="fn-bracket">[</span><a role="doc-backlink" href="#id1">1</a><span class="fn-bracket">]</span></span>
<p>Depending on the algorithm the mini-batch size can influence results or
not. SGD*, PassiveAggressive*, and discrete NaiveBayes are truly online
and are not affected by batch size. Conversely, MiniBatchKMeans
convergence rate is affected by the batch size. Also, its memory
footprint can vary dramatically with batch size.</p>
</aside>
</aside>
</section>
</section>
</section>
</div>
<div class="container">
<footer class="sk-content-footer">
© 2007 - 2022, scikit-learn developers (BSD License).
<a href="../_sources/computing/scaling_strategies.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>
$(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');
var hide_text = 'Hide prompts and outputs';
var show_text = 'Show prompts and outputs';
// create and add the button to all the code blocks that contain >>>
div.each(function(index) {
var jthis = $(this);
if (jthis.find('.gp').length > 0) {
var button = $('<span class="copybutton">>>></span>');
button.attr('title', hide_text);
button.data('hidden', 'false');
jthis.prepend(button);
}
// 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>');
});
// define the behavior of the button when it's clicked
$('.copybutton').click(function(e){
e.preventDefault();
var button = $(this);
if (button.data('hidden') === 'false') {
// hide the code output
button.parent().find('.go, .gp, .gt').hide();
button.next('pre').find('.gt').nextUntil('.gp, .go').css('visibility', 'hidden');
button.css('text-decoration', 'line-through');
button.attr('title', show_text);
button.data('hidden', 'true');
} else {
// show the code output
button.parent().find('.go, .gp, .gt').show();
button.next('pre').find('.gt').nextUntil('.gp, .go').css('visibility', 'visible');
button.css('text-decoration', 'none');
button.attr('title', hide_text);
button.data('hidden', 'false');
}
});
/*** 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>');
});
/*** Hide navbar when scrolling down ***/
// Returns true when headerlink target matches hash in url
(function() {
hashTargetOnTop = function() {
var hash = window.location.hash;
if ( hash.length < 2 ) { return false; }
var target = document.getElementById( hash.slice(1) );
if ( target === null ) { return false; }
var top = target.getBoundingClientRect().top;
return (top < 2) && (top > -2);
};
// Hide navbar on load if hash target is on top
var navBar = document.getElementById("navbar");
var navBarToggler = document.getElementById("sk-navbar-toggler");
var navBarHeightHidden = "-" + navBar.getBoundingClientRect().height + "px";
var $window = $(window);
hideNavBar = function() {
navBar.style.top = navBarHeightHidden;
};
showNavBar = function() {
navBar.style.top = "0";
}
if (hashTargetOnTop()) {
hideNavBar()
}
var prevScrollpos = window.pageYOffset;
hideOnScroll = function(lastScrollTop) {
if (($window.width() < 768) && (navBarToggler.getAttribute("aria-expanded") === 'true')) {
return;
}
if (lastScrollTop > 2 && (prevScrollpos <= lastScrollTop) || hashTargetOnTop()){
hideNavBar()
} else {
showNavBar()
}
prevScrollpos = lastScrollTop;
};
/*** high performance scroll event listener***/
var raf = window.requestAnimationFrame ||
window.webkitRequestAnimationFrame ||
window.mozRequestAnimationFrame ||
window.msRequestAnimationFrame ||
window.oRequestAnimationFrame;
var lastScrollTop = $window.scrollTop();
if (raf) {
loop();
}
function loop() {
var scrollTop = $window.scrollTop();
if (lastScrollTop === scrollTop) {
raf(loop);
return;
} else {
lastScrollTop = scrollTop;
hideOnScroll(lastScrollTop);
raf(loop);
}
}
})();
});
</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>