-
Notifications
You must be signed in to change notification settings - Fork 2
/
case-study-text-classification-spam-and-ham-.html
488 lines (436 loc) · 37.8 KB
/
case-study-text-classification-spam-and-ham-.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
<!DOCTYPE html>
<html >
<head>
<meta charset="UTF-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<title>Chapter 15 Case Study - Text classification: Spam and Ham. | Machine Learning with R</title>
<meta name="description" content="This book is about using R for machine learning purposes.">
<meta name="generator" content="bookdown and GitBook 2.6.7">
<meta property="og:title" content="Chapter 15 Case Study - Text classification: Spam and Ham. | Machine Learning with R" />
<meta property="og:type" content="book" />
<meta property="og:description" content="This book is about using R for machine learning purposes." />
<meta name="github-repo" content="fderyckel/machinelearningwithr" />
<meta name="twitter:card" content="summary" />
<meta name="twitter:title" content="Chapter 15 Case Study - Text classification: Spam and Ham. | Machine Learning with R" />
<meta name="twitter:description" content="This book is about using R for machine learning purposes." />
<meta name="author" content="François de Ryckel">
<meta name="date" content="2019-02-23">
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta name="apple-mobile-web-app-capable" content="yes">
<meta name="apple-mobile-web-app-status-bar-style" content="black">
<link rel="prev" href="model-evaluation.html">
<link rel="next" href="mushroom.html">
<script src="libs/jquery-2.2.3/jquery.min.js"></script>
<link href="libs/gitbook-2.6.7/css/style.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-table.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-bookdown.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-highlight.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-search.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-fontsettings.css" rel="stylesheet" />
<script src="libs/kePrint-0.0.1/kePrint.js"></script>
<style type="text/css">
div.sourceCode { overflow-x: auto; }
table.sourceCode, tr.sourceCode, td.lineNumbers, td.sourceCode {
margin: 0; padding: 0; vertical-align: baseline; border: none; }
table.sourceCode { width: 100%; line-height: 100%; }
td.lineNumbers { text-align: right; padding-right: 4px; padding-left: 4px; color: #aaaaaa; border-right: 1px solid #aaaaaa; }
td.sourceCode { padding-left: 5px; }
code > span.kw { color: #007020; font-weight: bold; } /* Keyword */
code > span.dt { color: #902000; } /* DataType */
code > span.dv { color: #40a070; } /* DecVal */
code > span.bn { color: #40a070; } /* BaseN */
code > span.fl { color: #40a070; } /* Float */
code > span.ch { color: #4070a0; } /* Char */
code > span.st { color: #4070a0; } /* String */
code > span.co { color: #60a0b0; font-style: italic; } /* Comment */
code > span.ot { color: #007020; } /* Other */
code > span.al { color: #ff0000; font-weight: bold; } /* Alert */
code > span.fu { color: #06287e; } /* Function */
code > span.er { color: #ff0000; font-weight: bold; } /* Error */
code > span.wa { color: #60a0b0; font-weight: bold; font-style: italic; } /* Warning */
code > span.cn { color: #880000; } /* Constant */
code > span.sc { color: #4070a0; } /* SpecialChar */
code > span.vs { color: #4070a0; } /* VerbatimString */
code > span.ss { color: #bb6688; } /* SpecialString */
code > span.im { } /* Import */
code > span.va { color: #19177c; } /* Variable */
code > span.cf { color: #007020; font-weight: bold; } /* ControlFlow */
code > span.op { color: #666666; } /* Operator */
code > span.bu { } /* BuiltIn */
code > span.ex { } /* Extension */
code > span.pp { color: #bc7a00; } /* Preprocessor */
code > span.at { color: #7d9029; } /* Attribute */
code > span.do { color: #ba2121; font-style: italic; } /* Documentation */
code > span.an { color: #60a0b0; font-weight: bold; font-style: italic; } /* Annotation */
code > span.cv { color: #60a0b0; font-weight: bold; font-style: italic; } /* CommentVar */
code > span.in { color: #60a0b0; font-weight: bold; font-style: italic; } /* Information */
</style>
<link rel="stylesheet" href="style.css" type="text/css" />
</head>
<body>
<div class="book without-animation with-summary font-size-2 font-family-1" data-basepath=".">
<div class="book-summary">
<nav role="navigation">
<ul class="summary">
<li><strong><a href="./">Machine Learning with R</a></strong></li>
<li class="divider"></li>
<li class="chapter" data-level="1" data-path="index.html"><a href="index.html"><i class="fa fa-check"></i><b>1</b> Prerequisites</a><ul>
<li class="chapter" data-level="1.1" data-path="index.html"><a href="index.html#pre-requisite-and-conventions"><i class="fa fa-check"></i><b>1.1</b> Pre-requisite and conventions</a></li>
<li class="chapter" data-level="1.2" data-path="index.html"><a href="index.html#organization"><i class="fa fa-check"></i><b>1.2</b> Organization</a></li>
<li class="chapter" data-level="1.3" data-path="index.html"><a href="index.html#packages"><i class="fa fa-check"></i><b>1.3</b> Packages</a></li>
</ul></li>
<li class="chapter" data-level="2" data-path="testinference.html"><a href="testinference.html"><i class="fa fa-check"></i><b>2</b> Tests and inferences</a><ul>
<li class="chapter" data-level="2.1" data-path="testinference.html"><a href="testinference.html#normality"><i class="fa fa-check"></i><b>2.1</b> Assumption of normality</a><ul>
<li class="chapter" data-level="2.1.1" data-path="testinference.html"><a href="testinference.html#visual-check-of-normality"><i class="fa fa-check"></i><b>2.1.1</b> Visual check of normality</a></li>
<li class="chapter" data-level="2.1.2" data-path="testinference.html"><a href="testinference.html#normality-tests"><i class="fa fa-check"></i><b>2.1.2</b> Normality tests</a></li>
</ul></li>
<li class="chapter" data-level="2.2" data-path="testinference.html"><a href="testinference.html#ttest"><i class="fa fa-check"></i><b>2.2</b> T-tests</a></li>
<li class="chapter" data-level="2.3" data-path="testinference.html"><a href="testinference.html#anova---analyse-of-variance."><i class="fa fa-check"></i><b>2.3</b> ANOVA - Analyse of variance.</a></li>
<li class="chapter" data-level="2.4" data-path="testinference.html"><a href="testinference.html#covariance"><i class="fa fa-check"></i><b>2.4</b> Covariance</a></li>
</ul></li>
<li class="chapter" data-level="3" data-path="mlr.html"><a href="mlr.html"><i class="fa fa-check"></i><b>3</b> Single & Multiple Linear Regression</a><ul>
<li class="chapter" data-level="3.1" data-path="mlr.html"><a href="mlr.html#single-variable-regression"><i class="fa fa-check"></i><b>3.1</b> Single variable regression</a></li>
<li class="chapter" data-level="3.2" data-path="mlr.html"><a href="mlr.html#multi-variables-regression"><i class="fa fa-check"></i><b>3.2</b> Multi-variables regression</a><ul>
<li class="chapter" data-level="3.2.1" data-path="mlr.html"><a href="mlr.html#predicting-wine-price-again"><i class="fa fa-check"></i><b>3.2.1</b> Predicting wine price (again!)</a></li>
</ul></li>
<li class="chapter" data-level="3.3" data-path="mlr.html"><a href="mlr.html#model-diagnostic-and-evaluation"><i class="fa fa-check"></i><b>3.3</b> Model diagnostic and evaluation</a></li>
<li class="chapter" data-level="3.4" data-path="mlr.html"><a href="mlr.html#final-example---boston-dataset---with-backward-elimination"><i class="fa fa-check"></i><b>3.4</b> Final example - Boston dataset - with backward elimination</a><ul>
<li class="chapter" data-level="3.4.1" data-path="mlr.html"><a href="mlr.html#model-diagmostic"><i class="fa fa-check"></i><b>3.4.1</b> Model diagmostic</a></li>
</ul></li>
<li class="chapter" data-level="3.5" data-path="mlr.html"><a href="mlr.html#references"><i class="fa fa-check"></i><b>3.5</b> References</a></li>
</ul></li>
<li class="chapter" data-level="4" data-path="logistic.html"><a href="logistic.html"><i class="fa fa-check"></i><b>4</b> Logistic Regression</a><ul>
<li class="chapter" data-level="4.1" data-path="logistic.html"><a href="logistic.html#introduction"><i class="fa fa-check"></i><b>4.1</b> Introduction</a></li>
<li class="chapter" data-level="4.2" data-path="logistic.html"><a href="logistic.html#the-logistic-equation."><i class="fa fa-check"></i><b>4.2</b> The logistic equation.</a></li>
<li class="chapter" data-level="4.3" data-path="logistic.html"><a href="logistic.html#performance-of-logistic-regression-model"><i class="fa fa-check"></i><b>4.3</b> Performance of Logistic Regression Model</a></li>
<li class="chapter" data-level="4.4" data-path="logistic.html"><a href="logistic.html#setting-up"><i class="fa fa-check"></i><b>4.4</b> Setting up</a></li>
<li class="chapter" data-level="4.5" data-path="logistic.html"><a href="logistic.html#example-1---graduate-admission"><i class="fa fa-check"></i><b>4.5</b> Example 1 - Graduate Admission</a></li>
<li class="chapter" data-level="4.6" data-path="logistic.html"><a href="logistic.html#example-2---diabetes"><i class="fa fa-check"></i><b>4.6</b> Example 2 - Diabetes</a><ul>
<li class="chapter" data-level="4.6.1" data-path="logistic.html"><a href="logistic.html#accounting-for-missing-values"><i class="fa fa-check"></i><b>4.6.1</b> Accounting for missing values</a></li>
<li class="chapter" data-level="4.6.2" data-path="logistic.html"><a href="logistic.html#imputting-missing-values"><i class="fa fa-check"></i><b>4.6.2</b> Imputting Missing Values</a></li>
<li class="chapter" data-level="4.6.3" data-path="logistic.html"><a href="logistic.html#roc-and-auc"><i class="fa fa-check"></i><b>4.6.3</b> ROC and AUC</a></li>
</ul></li>
<li class="chapter" data-level="4.7" data-path="logistic.html"><a href="logistic.html#references-1"><i class="fa fa-check"></i><b>4.7</b> References</a></li>
</ul></li>
<li class="chapter" data-level="5" data-path="softmax-and-multinomial-regressions.html"><a href="softmax-and-multinomial-regressions.html"><i class="fa fa-check"></i><b>5</b> Softmax and multinomial regressions</a><ul>
<li class="chapter" data-level="5.1" data-path="softmax-and-multinomial-regressions.html"><a href="softmax-and-multinomial-regressions.html#multinomial-logistic-regression"><i class="fa fa-check"></i><b>5.1</b> Multinomial Logistic Regression</a></li>
<li class="chapter" data-level="5.2" data-path="softmax-and-multinomial-regressions.html"><a href="softmax-and-multinomial-regressions.html#references-2"><i class="fa fa-check"></i><b>5.2</b> References</a></li>
</ul></li>
<li class="chapter" data-level="6" data-path="gradient-descent.html"><a href="gradient-descent.html"><i class="fa fa-check"></i><b>6</b> Gradient Descent</a><ul>
<li class="chapter" data-level="6.1" data-path="gradient-descent.html"><a href="gradient-descent.html#example-on-functions"><i class="fa fa-check"></i><b>6.1</b> Example on functions</a></li>
<li class="chapter" data-level="6.2" data-path="gradient-descent.html"><a href="gradient-descent.html#example-on-regressions"><i class="fa fa-check"></i><b>6.2</b> Example on regressions</a></li>
</ul></li>
<li class="chapter" data-level="7" data-path="knnchapter.html"><a href="knnchapter.html"><i class="fa fa-check"></i><b>7</b> KNN - K Nearest Neighbour</a><ul>
<li class="chapter" data-level="7.1" data-path="knnchapter.html"><a href="knnchapter.html#example-1.-prostate-cancer-dataset"><i class="fa fa-check"></i><b>7.1</b> Example 1. Prostate Cancer dataset</a></li>
<li class="chapter" data-level="7.2" data-path="knnchapter.html"><a href="knnchapter.html#example-2.-wine-dataset"><i class="fa fa-check"></i><b>7.2</b> Example 2. Wine dataset</a><ul>
<li class="chapter" data-level="7.2.1" data-path="knnchapter.html"><a href="knnchapter.html#understand-the-data"><i class="fa fa-check"></i><b>7.2.1</b> Understand the data</a></li>
</ul></li>
<li class="chapter" data-level="7.3" data-path="knnchapter.html"><a href="knnchapter.html#references-3"><i class="fa fa-check"></i><b>7.3</b> References</a></li>
</ul></li>
<li class="chapter" data-level="8" data-path="kmeans.html"><a href="kmeans.html"><i class="fa fa-check"></i><b>8</b> Kmeans clustering</a><ul>
<li class="chapter" data-level="8.1" data-path="kmeans.html"><a href="kmeans.html#multinomial-logistic-regression-1"><i class="fa fa-check"></i><b>8.1</b> Multinomial Logistic Regression</a></li>
<li class="chapter" data-level="8.2" data-path="kmeans.html"><a href="kmeans.html#references-4"><i class="fa fa-check"></i><b>8.2</b> References</a></li>
</ul></li>
<li class="chapter" data-level="9" data-path="hierclust.html"><a href="hierclust.html"><i class="fa fa-check"></i><b>9</b> Hierarichal Clustering</a><ul>
<li class="chapter" data-level="9.1" data-path="hierclust.html"><a href="hierclust.html#example-on-the-pokemon-dataset"><i class="fa fa-check"></i><b>9.1</b> Example on the Pokemon dataset</a></li>
<li class="chapter" data-level="9.2" data-path="hierclust.html"><a href="hierclust.html#example-on-regressions-1"><i class="fa fa-check"></i><b>9.2</b> Example on regressions</a></li>
<li class="chapter" data-level="9.3" data-path="hierclust.html"><a href="hierclust.html#references-5"><i class="fa fa-check"></i><b>9.3</b> References</a></li>
</ul></li>
<li class="chapter" data-level="10" data-path="pca.html"><a href="pca.html"><i class="fa fa-check"></i><b>10</b> Principal Component Analysis</a><ul>
<li class="chapter" data-level="10.1" data-path="pca.html"><a href="pca.html#pca-on-an-easy-example."><i class="fa fa-check"></i><b>10.1</b> PCA on an easy example.</a></li>
<li class="chapter" data-level="10.2" data-path="pca.html"><a href="pca.html#references."><i class="fa fa-check"></i><b>10.2</b> References.</a></li>
</ul></li>
<li class="chapter" data-level="11" data-path="trees-and-classification.html"><a href="trees-and-classification.html"><i class="fa fa-check"></i><b>11</b> Trees and Classification</a><ul>
<li class="chapter" data-level="11.1" data-path="trees-and-classification.html"><a href="trees-and-classification.html#introduction-1"><i class="fa fa-check"></i><b>11.1</b> Introduction</a></li>
<li class="chapter" data-level="11.2" data-path="trees-and-classification.html"><a href="trees-and-classification.html#first-example."><i class="fa fa-check"></i><b>11.2</b> First example.</a></li>
<li class="chapter" data-level="11.3" data-path="trees-and-classification.html"><a href="trees-and-classification.html#second-example."><i class="fa fa-check"></i><b>11.3</b> Second Example.</a></li>
<li class="chapter" data-level="11.4" data-path="trees-and-classification.html"><a href="trees-and-classification.html#how-does-a-tree-decide-where-to-split"><i class="fa fa-check"></i><b>11.4</b> How does a tree decide where to split?</a></li>
<li class="chapter" data-level="11.5" data-path="trees-and-classification.html"><a href="trees-and-classification.html#third-example."><i class="fa fa-check"></i><b>11.5</b> Third example.</a></li>
<li class="chapter" data-level="11.6" data-path="trees-and-classification.html"><a href="trees-and-classification.html#references-6"><i class="fa fa-check"></i><b>11.6</b> References</a></li>
</ul></li>
<li class="chapter" data-level="12" data-path="random-forest.html"><a href="random-forest.html"><i class="fa fa-check"></i><b>12</b> Random Forest</a><ul>
<li class="chapter" data-level="12.1" data-path="random-forest.html"><a href="random-forest.html#how-does-it-work"><i class="fa fa-check"></i><b>12.1</b> How does it work?</a></li>
<li class="chapter" data-level="12.2" data-path="random-forest.html"><a href="random-forest.html#references-7"><i class="fa fa-check"></i><b>12.2</b> References</a></li>
</ul></li>
<li class="chapter" data-level="13" data-path="svm.html"><a href="svm.html"><i class="fa fa-check"></i><b>13</b> Support Vector Machine</a><ul>
<li class="chapter" data-level="13.1" data-path="svm.html"><a href="svm.html#support-vecotr-regression"><i class="fa fa-check"></i><b>13.1</b> Support Vecotr Regression</a><ul>
<li class="chapter" data-level="13.1.1" data-path="svm.html"><a href="svm.html#create-data"><i class="fa fa-check"></i><b>13.1.1</b> Create data</a></li>
<li class="chapter" data-level="13.1.2" data-path="svm.html"><a href="svm.html#tuning-a-svm-model"><i class="fa fa-check"></i><b>13.1.2</b> Tuning a SVM model</a></li>
<li class="chapter" data-level="13.1.3" data-path="svm.html"><a href="svm.html#discussion-on-parameters"><i class="fa fa-check"></i><b>13.1.3</b> Discussion on parameters</a></li>
</ul></li>
<li class="chapter" data-level="13.2" data-path="svm.html"><a href="svm.html#references-8"><i class="fa fa-check"></i><b>13.2</b> References</a></li>
</ul></li>
<li class="chapter" data-level="14" data-path="model-evaluation.html"><a href="model-evaluation.html"><i class="fa fa-check"></i><b>14</b> Model Evaluation</a><ul>
<li class="chapter" data-level="14.1" data-path="model-evaluation.html"><a href="model-evaluation.html#biais-variance-tradeoff"><i class="fa fa-check"></i><b>14.1</b> Biais variance tradeoff</a></li>
<li class="chapter" data-level="14.2" data-path="model-evaluation.html"><a href="model-evaluation.html#bagging"><i class="fa fa-check"></i><b>14.2</b> Bagging</a></li>
<li class="chapter" data-level="14.3" data-path="model-evaluation.html"><a href="model-evaluation.html#crossvalidation"><i class="fa fa-check"></i><b>14.3</b> Cross Validation</a></li>
</ul></li>
<li class="chapter" data-level="15" data-path="case-study-text-classification-spam-and-ham-.html"><a href="case-study-text-classification-spam-and-ham-.html"><i class="fa fa-check"></i><b>15</b> Case Study - Text classification: Spam and Ham.</a></li>
<li class="chapter" data-level="16" data-path="mushroom.html"><a href="mushroom.html"><i class="fa fa-check"></i><b>16</b> Case Study - Mushrooms Classification</a><ul>
<li class="chapter" data-level="16.1" data-path="mushroom.html"><a href="mushroom.html#import-the-data"><i class="fa fa-check"></i><b>16.1</b> Import the data</a></li>
<li class="chapter" data-level="16.2" data-path="mushroom.html"><a href="mushroom.html#tidy-the-data"><i class="fa fa-check"></i><b>16.2</b> Tidy the data</a></li>
<li class="chapter" data-level="16.3" data-path="mushroom.html"><a href="mushroom.html#understand-the-data-1"><i class="fa fa-check"></i><b>16.3</b> Understand the data</a><ul>
<li class="chapter" data-level="16.3.1" data-path="mushroom.html"><a href="mushroom.html#transform-the-data"><i class="fa fa-check"></i><b>16.3.1</b> Transform the data</a></li>
<li class="chapter" data-level="16.3.2" data-path="mushroom.html"><a href="mushroom.html#visualize-the-data"><i class="fa fa-check"></i><b>16.3.2</b> Visualize the data</a></li>
<li class="chapter" data-level="16.3.3" data-path="mushroom.html"><a href="mushroom.html#modeling"><i class="fa fa-check"></i><b>16.3.3</b> Modeling</a></li>
</ul></li>
<li class="chapter" data-level="16.4" data-path="mushroom.html"><a href="mushroom.html#communication"><i class="fa fa-check"></i><b>16.4</b> Communication</a></li>
</ul></li>
<li class="chapter" data-level="17" data-path="case-study-the-adults-dataset-.html"><a href="case-study-the-adults-dataset-.html"><i class="fa fa-check"></i><b>17</b> Case study - The adults dataset.</a><ul>
<li class="chapter" data-level="17.1" data-path="case-study-the-adults-dataset-.html"><a href="case-study-the-adults-dataset-.html#introduction-2"><i class="fa fa-check"></i><b>17.1</b> Introduction</a></li>
<li class="chapter" data-level="17.2" data-path="case-study-the-adults-dataset-.html"><a href="case-study-the-adults-dataset-.html#import-the-data-1"><i class="fa fa-check"></i><b>17.2</b> Import the data</a></li>
<li class="chapter" data-level="17.3" data-path="case-study-the-adults-dataset-.html"><a href="case-study-the-adults-dataset-.html#tidy-the-data-1"><i class="fa fa-check"></i><b>17.3</b> Tidy the data</a></li>
</ul></li>
<li class="chapter" data-level="18" data-path="breastcancer.html"><a href="breastcancer.html"><i class="fa fa-check"></i><b>18</b> Case Study - Wisconsin Breast Cancer</a><ul>
<li class="chapter" data-level="18.1" data-path="breastcancer.html"><a href="breastcancer.html#import-the-data-2"><i class="fa fa-check"></i><b>18.1</b> Import the data</a></li>
<li class="chapter" data-level="18.2" data-path="breastcancer.html"><a href="breastcancer.html#tidy-the-data-2"><i class="fa fa-check"></i><b>18.2</b> Tidy the data</a></li>
<li class="chapter" data-level="18.3" data-path="breastcancer.html"><a href="breastcancer.html#understand-the-data-2"><i class="fa fa-check"></i><b>18.3</b> Understand the data</a><ul>
<li class="chapter" data-level="18.3.1" data-path="breastcancer.html"><a href="breastcancer.html#transform-the-data-1"><i class="fa fa-check"></i><b>18.3.1</b> Transform the data</a></li>
<li class="chapter" data-level="18.3.2" data-path="breastcancer.html"><a href="breastcancer.html#pre-process-the-data"><i class="fa fa-check"></i><b>18.3.2</b> Pre-process the data</a></li>
<li class="chapter" data-level="18.3.3" data-path="breastcancer.html"><a href="breastcancer.html#model-the-data-1"><i class="fa fa-check"></i><b>18.3.3</b> Model the data</a></li>
</ul></li>
<li class="chapter" data-level="18.4" data-path="breastcancer.html"><a href="breastcancer.html#references-9"><i class="fa fa-check"></i><b>18.4</b> References</a></li>
</ul></li>
<li class="chapter" data-level="19" data-path="final-words.html"><a href="final-words.html"><i class="fa fa-check"></i><b>19</b> Final Words</a></li>
<li class="chapter" data-level="" data-path="references-10.html"><a href="references-10.html"><i class="fa fa-check"></i>References</a></li>
</ul>
</nav>
</div>
<div class="book-body">
<div class="body-inner">
<div class="book-header" role="navigation">
<h1>
<i class="fa fa-circle-o-notch fa-spin"></i><a href="./">Machine Learning with R</a>
</h1>
</div>
<div class="page-wrapper" tabindex="-1" role="main">
<div class="page-inner">
<section class="normal" id="section-">
<div id="case-study---text-classification-spam-and-ham." class="section level1">
<h1><span class="header-section-number">Chapter 15</span> Case Study - Text classification: Spam and Ham.</h1>
<p>This chapter has been inspired by the Coursera course on <a href="https://www.coursera.org/learn/ml-foundations/">Machine Learning Foundations: A Case Study Approach</a> given by Carlos Guestrin and by Emily Fox from Washington University. This course is part of the <a href="https://www.coursera.org/specializations/machine-learning">Machine Learning Specialization</a></p>
<p>The task was to apply classfification on an Amazon review dataset. Given a review, we create a model that will decide if the review is <em>positive</em> (associated with a rating of 4 or 5) or <em>negative</em> (associate with a rating of 1 or 2). This is a supervised learning task as the grading associated with the reviews is used as the response variable.</p>
<p>What we have done here is to create a subset of the dataset with only one product. The <em>Philips Avent Bottle</em>.</p>
<p>As usual, let’s first load the libraries</p>
<p>Let’s have a quick look at our data.</p>
<div class="sourceCode"><pre class="sourceCode r"><code class="sourceCode r">df <-<span class="st"> </span><span class="kw">read_csv</span>(<span class="st">"dataset/toyamazonPhilips.csv"</span>)
df <-<span class="st"> </span><span class="kw">as_tibble</span>(df)
df2 <-<span class="st"> </span>df[,<span class="dv">2</span><span class="op">:</span><span class="dv">3</span>]
<span class="co">#Let's have a quick look at the reviews</span>
<span class="kw">library</span>(pander)
<span class="kw">pandoc.table</span>(df2[<span class="dv">1</span><span class="op">:</span><span class="dv">3</span>,],
<span class="dt">justify =</span> <span class="st">'left'</span>, <span class="dt">style =</span> <span class="st">'grid'</span>)</code></pre></div>
<pre><code>##
##
## +--------------------------------+--------+
## | review | rating |
## +================================+========+
## | I was recommended to use these | 5 |
## | bottles by a girlfriend who | |
## | had used them with her baby | |
## | and sworn by them. I started | |
## | with a set of 4oz bottles with | |
## | newborn nipples (#1) and have | |
## | graduated now to the 9oz | |
## | bottles with the next size up | |
## | nipples (#2). I will simply | |
## | buy new nipples when she's | |
## | ready for the next step. I am | |
## | absolutely thrilled with these | |
## | bottles. I do have an | |
## | occasional leak but it is | |
## | always my own fault for not | |
## | screwing the lid on tight | |
## | enough!I have not had any | |
## | problems with leaking when I | |
## | have them put together | |
## | correctly. They were | |
## | especially good for going back | |
## | and forth between breast | |
## | feeding and bottle feeding. My | |
## | now 4mos old baby never had | |
## | any difficulties feeding | |
## | either way and transitioned | |
## | very smoothly. One reader | |
## | referred to air being pushed | |
## | into the bottle during feeding | |
## | and causing her baby to | |
## | swallow air. The air wooshing | |
## | in is a good thing! It | |
## | prevents a vacuum from | |
## | developing inside the bottle | |
## | and allows the baby to | |
## | continuously feed without | |
## | having to stop to relieve the | |
## | pressure inside the bottle. | |
## | The only time I ever had any | |
## | problems with mine swallowing | |
## | air was when she had pretty | |
## | much outgrown the newborn | |
## | nipples and would try to suck | |
## | too hard when she was very | |
## | hungry. This was immediately | |
## | corrected by buying the next | |
## | size up. They're also very | |
## | easy to clean and can be | |
## | effectively washed in the | |
## | dishwasher because of their | |
## | wide neck. Regular bottles | |
## | cannot. I, out of necessity, | |
## | used another brand of bottle | |
## | with her today at a relative's | |
## | house and was reminded why I | |
## | like Avent bottles so much! | |
## +--------------------------------+--------+
## | If I had not been given a ton | 2 |
## | of Avent bottles, I would have | |
## | chosen some other system. The | |
## | leaking is terrible!!! You | |
## | have to buy the disks | |
## | separately, you should get | |
## | them for free because they are | |
## | absolutely essential. The | |
## | only way to mix formula in the | |
## | bottle or transport liquid is | |
## | to use the disks in the ring, | |
## | then switch to the nipple when | |
## | you are ready to feed. The | |
## | only reason I give it a two is | |
## | because I do like that you can | |
## | pump directly into the bottle | |
## | with the ISIS breast pump. | |
## | And, I like the sippy cups. | |
## +--------------------------------+--------+
## | Leaks! Especially difficult to | 1 |
## | get a tight seal if you use | |
## | one hand (while holding baby). | |
## | A much better design is the | |
## | Breast Flow Learning Curve | |
## | First Years bottles. Instead | |
## | buy The First Years 3pk. | |
## | Breastflow 5oz. Bottles These | |
## | worked much better for me. | |
## +--------------------------------+--------+</code></pre>
<div class="sourceCode"><pre class="sourceCode r"><code class="sourceCode r"><span class="co">#Let's see the table of ratings. </span>
<span class="kw">table</span>(df2<span class="op">$</span>rating)</code></pre></div>
<pre><code>##
## 1 2 3 4 5
## 45 33 17 30 66</code></pre>
<p>Interestingly the ratings on the Avent Bottles are quite spread on the extreme. It might be that people only write reviews if they are super excited or very frustrated with a product. Because we want this to be a binary classification exercise, we’ll do some transformation on these ratings. First we combine the positive reviews together (the 4 and 5 ratings) and the negative reviews together (the 1 and 2 ratings). Then we take out the neutral reviews.</p>
<div class="sourceCode"><pre class="sourceCode r"><code class="sourceCode r"><span class="co"># We'll put a 1 for great reviews (4 or 5) or a 0 for bad reviews (1 or 2)</span>
<span class="co"># We remove all the reviews that have a rating of 3</span>
df2 <-<span class="st"> </span>df <span class="op">%>%</span><span class="st"> </span><span class="kw">filter</span>(rating <span class="op">!=</span><span class="st"> </span><span class="dv">3</span>) <span class="op">%>%</span><span class="st"> </span>
<span class="st"> </span><span class="kw">mutate</span>(<span class="dt">rating_new =</span> <span class="kw">if_else</span>(rating <span class="op">>=</span><span class="st"> </span><span class="dv">4</span>, <span class="dv">1</span>, <span class="dv">0</span>))
df_training <-<span class="st"> </span>df2[<span class="dv">1</span><span class="op">:</span><span class="dv">150</span>, ]</code></pre></div>
<p>Now we create our corpus, then tokenize it, then make it back to a data frame.</p>
<div class="sourceCode"><pre class="sourceCode r"><code class="sourceCode r"><span class="kw">library</span>(tm)
corpus_toy <-<span class="st"> </span><span class="kw">Corpus</span>(<span class="kw">VectorSource</span>(df_training<span class="op">$</span>review))
tdm_toy <-<span class="st"> </span><span class="kw">DocumentTermMatrix</span>(corpus_toy, <span class="kw">list</span>(<span class="dt">removePunctuation =</span> <span class="ot">TRUE</span>,
<span class="dt">removeNumbers =</span> <span class="ot">TRUE</span>))
training_set_toy <-<span class="st"> </span><span class="kw">as.matrix</span>(tdm_toy)
training_set_toy <-<span class="st"> </span><span class="kw">cbind</span>(training_set_toy, df_training<span class="op">$</span>rating_new)
<span class="kw">colnames</span>(training_set_toy)[<span class="kw">ncol</span>(training_set_toy)] <-<span class="st"> "y"</span>
training_set_toy <-<span class="st"> </span><span class="kw">as.data.frame</span>(training_set_toy)
training_set_toy<span class="op">$</span>y <-<span class="st"> </span><span class="kw">as.factor</span>(training_set_toy<span class="op">$</span>y)</code></pre></div>
<p>Now that we have our data frame ready, let’s create our model using the <code>svmLinear3</code> method.</p>
<div class="sourceCode"><pre class="sourceCode r"><code class="sourceCode r">review_toy_model <-<span class="st"> </span>caret<span class="op">::</span><span class="kw">train</span>(y <span class="op">~</span>., <span class="dt">data =</span> training_set_toy, <span class="dt">method =</span> <span class="st">'svmLinear3'</span>)</code></pre></div>
<p>Now we try our model on new review data</p>
<div class="sourceCode"><pre class="sourceCode r"><code class="sourceCode r">test_review_data <-<span class="st"> </span>df2[<span class="dv">151</span><span class="op">:</span><span class="dv">174</span>, ]
test_corpus <-<span class="st"> </span><span class="kw">Corpus</span>(<span class="kw">VectorSource</span>(test_review_data<span class="op">$</span>review))
test_tdm <-<span class="st"> </span><span class="kw">DocumentTermMatrix</span>(test_corpus, <span class="dt">control=</span><span class="kw">list</span>(<span class="dt">dictionary =</span> <span class="kw">Terms</span>(tdm_toy)))
test_tdm <-<span class="st"> </span><span class="kw">as.matrix</span>(test_tdm)
<span class="co">#Build the prediction </span>
model_toy_result <-<span class="st"> </span><span class="kw">predict</span>(review_toy_model, <span class="dt">newdata =</span> test_tdm)
check_accuracy <-<span class="st"> </span><span class="kw">as.data.frame</span>(<span class="kw">cbind</span>(<span class="dt">prediction =</span> model_toy_result,
<span class="dt">rating =</span> test_review_data<span class="op">$</span>rating_new))
check_accuracy <-<span class="st"> </span>check_accuracy <span class="op">%>%</span><span class="st"> </span><span class="kw">mutate</span>(<span class="dt">prediction =</span> <span class="kw">as.integer</span>(prediction) <span class="op">-</span><span class="st"> </span><span class="dv">1</span>)
check_accuracy<span class="op">$</span>accuracy <-<span class="st"> </span><span class="kw">if_else</span>(check_accuracy<span class="op">$</span>prediction <span class="op">==</span><span class="st"> </span>check_accuracy<span class="op">$</span>rating, <span class="dv">1</span>, <span class="dv">0</span>)
<span class="kw">round</span>(<span class="kw">prop.table</span>(<span class="kw">table</span>(check_accuracy<span class="op">$</span>accuracy)), <span class="dv">3</span>)</code></pre></div>
<pre><code>##
## 0 1
## 0.208 0.792</code></pre>
<p>Another way to deal with text classification is to use the <code>RtextTool</code> library.<br />
We can use the same dataframe that we used in our previous method. Like before we “DocumentTermMatrix”, we create a matrix of terms</p>
<div class="sourceCode"><pre class="sourceCode r"><code class="sourceCode r"><span class="kw">library</span>(RTextTools)
product_review_matrix <-<span class="st"> </span><span class="kw">create_matrix</span>(df2[,<span class="dv">2</span>], <span class="dt">language =</span> <span class="st">"English"</span>,
<span class="dt">removeNumbers =</span> <span class="ot">TRUE</span>,
<span class="dt">removePunctuation =</span> <span class="ot">TRUE</span>,
<span class="dt">removeStopwords =</span> <span class="ot">FALSE</span>, <span class="dt">stemWords =</span> <span class="ot">FALSE</span>)
product_review_container <-<span class="st"> </span><span class="kw">create_container</span>(product_review_matrix,
df2<span class="op">$</span>rating_new,
<span class="dt">trainSize =</span> <span class="dv">1</span><span class="op">:</span><span class="dv">150</span>, <span class="dt">testSize =</span> <span class="dv">151</span><span class="op">:</span><span class="dv">174</span>,
<span class="dt">virgin =</span> <span class="ot">FALSE</span>)
product_review_model <-<span class="st"> </span><span class="kw">train_model</span>(product_review_container, <span class="dt">algorithm =</span> <span class="st">"SVM"</span>)
product_review_model_result <-<span class="st"> </span><span class="kw">classify_model</span>(product_review_container, product_review_model)
x <-<span class="st"> </span><span class="kw">as.data.frame</span>(<span class="kw">cbind</span>(df2<span class="op">$</span>rating_new[<span class="dv">151</span><span class="op">:</span><span class="dv">174</span>], product_review_model_result<span class="op">$</span>SVM_LABEL))
<span class="kw">colnames</span>(x) <-<span class="st"> </span><span class="kw">c</span>(<span class="st">"actual_ratings"</span>, <span class="st">"predicted_ratings"</span>)
x <-<span class="st"> </span>x <span class="op">%>%</span><span class="st"> </span><span class="kw">mutate</span>(<span class="dt">predicted_ratings =</span> predicted_ratings <span class="op">-</span><span class="st"> </span><span class="dv">1</span>)
<span class="kw">round</span>(<span class="kw">prop.table</span>(<span class="kw">table</span>(x<span class="op">$</span>actual_ratings <span class="op">==</span><span class="st"> </span>x<span class="op">$</span>predicted_ratings)), <span class="dv">3</span>)</code></pre></div>
<pre><code>##
## FALSE TRUE
## 0.25 0.75</code></pre>
</div>
</section>
</div>
</div>
</div>
<a href="model-evaluation.html" class="navigation navigation-prev " aria-label="Previous page"><i class="fa fa-angle-left"></i></a>
<a href="mushroom.html" class="navigation navigation-next " aria-label="Next page"><i class="fa fa-angle-right"></i></a>
</div>
</div>
<script src="libs/gitbook-2.6.7/js/app.min.js"></script>
<script src="libs/gitbook-2.6.7/js/lunr.js"></script>
<script src="libs/gitbook-2.6.7/js/plugin-search.js"></script>
<script src="libs/gitbook-2.6.7/js/plugin-sharing.js"></script>
<script src="libs/gitbook-2.6.7/js/plugin-fontsettings.js"></script>
<script src="libs/gitbook-2.6.7/js/plugin-bookdown.js"></script>
<script src="libs/gitbook-2.6.7/js/jquery.highlight.js"></script>
<script>
gitbook.require(["gitbook"], function(gitbook) {
gitbook.start({
"sharing": {
"github": false,
"facebook": true,
"twitter": true,
"google": false,
"linkedin": false,
"weibo": false,
"instapaper": false,
"vk": false,
"all": ["facebook", "google", "twitter", "linkedin", "weibo", "instapaper"]
},
"fontsettings": {
"theme": "white",
"family": "sans",
"size": 2
},
"edit": {
"link": "https://github.com/fderyckel/machinelearningwithr/edit/master/23-spam_and_ham.Rmd",
"text": "Suggest edit to this page"
},
"history": {
"link": null,
"text": null
},
"download": null,
"toc": {
"collapse": "section"
}
});
});
</script>
<!-- dynamically load mathjax for compatibility with self-contained -->
<script>
(function () {
var script = document.createElement("script");
script.type = "text/javascript";
var src = "";
if (src === "" || src === "true") src = "https://mathjax.rstudio.com/latest/MathJax.js?config=TeX-MML-AM_CHTML";
if (location.protocol !== "file:" && /^https?:/.test(src))
src = src.replace(/^https?:/, '');
script.src = src;
document.getElementsByTagName("head")[0].appendChild(script);
})();
</script>
</body>
</html>