-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathreadme.html
460 lines (390 loc) · 14.5 KB
/
readme.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
<!DOCTYPE html>
<html>
<head>
<title>README.md</title>
<meta http-equiv="Content-type" content="text/html;charset=UTF-8">
<style>
/* https://github.com/microsoft/vscode/blob/master/extensions/markdown-language-features/media/markdown.css */
/*---------------------------------------------------------------------------------------------
* Copyright (c) Microsoft Corporation. All rights reserved.
* Licensed under the MIT License. See License.txt in the project root for license information.
*--------------------------------------------------------------------------------------------*/
body {
font-family: var(--vscode-markdown-font-family, -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", "Ubuntu", "Droid Sans", sans-serif);
font-size: var(--vscode-markdown-font-size, 14px);
padding: 0 26px;
line-height: var(--vscode-markdown-line-height, 22px);
word-wrap: break-word;
}
#code-csp-warning {
position: fixed;
top: 0;
right: 0;
color: white;
margin: 16px;
text-align: center;
font-size: 12px;
font-family: sans-serif;
background-color:#444444;
cursor: pointer;
padding: 6px;
box-shadow: 1px 1px 1px rgba(0,0,0,.25);
}
#code-csp-warning:hover {
text-decoration: none;
background-color:#007acc;
box-shadow: 2px 2px 2px rgba(0,0,0,.25);
}
body.scrollBeyondLastLine {
margin-bottom: calc(100vh - 22px);
}
body.showEditorSelection .code-line {
position: relative;
}
body.showEditorSelection .code-active-line:before,
body.showEditorSelection .code-line:hover:before {
content: "";
display: block;
position: absolute;
top: 0;
left: -12px;
height: 100%;
}
body.showEditorSelection li.code-active-line:before,
body.showEditorSelection li.code-line:hover:before {
left: -30px;
}
.vscode-light.showEditorSelection .code-active-line:before {
border-left: 3px solid rgba(0, 0, 0, 0.15);
}
.vscode-light.showEditorSelection .code-line:hover:before {
border-left: 3px solid rgba(0, 0, 0, 0.40);
}
.vscode-light.showEditorSelection .code-line .code-line:hover:before {
border-left: none;
}
.vscode-dark.showEditorSelection .code-active-line:before {
border-left: 3px solid rgba(255, 255, 255, 0.4);
}
.vscode-dark.showEditorSelection .code-line:hover:before {
border-left: 3px solid rgba(255, 255, 255, 0.60);
}
.vscode-dark.showEditorSelection .code-line .code-line:hover:before {
border-left: none;
}
.vscode-high-contrast.showEditorSelection .code-active-line:before {
border-left: 3px solid rgba(255, 160, 0, 0.7);
}
.vscode-high-contrast.showEditorSelection .code-line:hover:before {
border-left: 3px solid rgba(255, 160, 0, 1);
}
.vscode-high-contrast.showEditorSelection .code-line .code-line:hover:before {
border-left: none;
}
img {
max-width: 100%;
max-height: 100%;
}
a {
text-decoration: none;
}
a:hover {
text-decoration: underline;
}
a:focus,
input:focus,
select:focus,
textarea:focus {
outline: 1px solid -webkit-focus-ring-color;
outline-offset: -1px;
}
hr {
border: 0;
height: 2px;
border-bottom: 2px solid;
}
h1 {
padding-bottom: 0.3em;
line-height: 1.2;
border-bottom-width: 1px;
border-bottom-style: solid;
}
h1, h2, h3 {
font-weight: normal;
}
table {
border-collapse: collapse;
}
table > thead > tr > th {
text-align: left;
border-bottom: 1px solid;
}
table > thead > tr > th,
table > thead > tr > td,
table > tbody > tr > th,
table > tbody > tr > td {
padding: 5px 10px;
}
table > tbody > tr + tr > td {
border-top: 1px solid;
}
blockquote {
margin: 0 7px 0 5px;
padding: 0 16px 0 10px;
border-left-width: 5px;
border-left-style: solid;
}
code {
font-family: Menlo, Monaco, Consolas, "Droid Sans Mono", "Courier New", monospace, "Droid Sans Fallback";
font-size: 1em;
line-height: 1.357em;
}
body.wordWrap pre {
white-space: pre-wrap;
}
pre:not(.hljs),
pre.hljs code > div {
padding: 16px;
border-radius: 3px;
overflow: auto;
}
pre code {
color: var(--vscode-editor-foreground);
tab-size: 4;
}
/** Theming */
.vscode-light pre {
background-color: rgba(220, 220, 220, 0.4);
}
.vscode-dark pre {
background-color: rgba(10, 10, 10, 0.4);
}
.vscode-high-contrast pre {
background-color: rgb(0, 0, 0);
}
.vscode-high-contrast h1 {
border-color: rgb(0, 0, 0);
}
.vscode-light table > thead > tr > th {
border-color: rgba(0, 0, 0, 0.69);
}
.vscode-dark table > thead > tr > th {
border-color: rgba(255, 255, 255, 0.69);
}
.vscode-light h1,
.vscode-light hr,
.vscode-light table > tbody > tr + tr > td {
border-color: rgba(0, 0, 0, 0.18);
}
.vscode-dark h1,
.vscode-dark hr,
.vscode-dark table > tbody > tr + tr > td {
border-color: rgba(255, 255, 255, 0.18);
}
</style>
<style>
/* Tomorrow Theme */
/* http://jmblog.github.com/color-themes-for-google-code-highlightjs */
/* Original theme - https://github.com/chriskempson/tomorrow-theme */
/* Tomorrow Comment */
.hljs-comment,
.hljs-quote {
color: #8e908c;
}
/* Tomorrow Red */
.hljs-variable,
.hljs-template-variable,
.hljs-tag,
.hljs-name,
.hljs-selector-id,
.hljs-selector-class,
.hljs-regexp,
.hljs-deletion {
color: #c82829;
}
/* Tomorrow Orange */
.hljs-number,
.hljs-built_in,
.hljs-builtin-name,
.hljs-literal,
.hljs-type,
.hljs-params,
.hljs-meta,
.hljs-link {
color: #f5871f;
}
/* Tomorrow Yellow */
.hljs-attribute {
color: #eab700;
}
/* Tomorrow Green */
.hljs-string,
.hljs-symbol,
.hljs-bullet,
.hljs-addition {
color: #718c00;
}
/* Tomorrow Blue */
.hljs-title,
.hljs-section {
color: #4271ae;
}
/* Tomorrow Purple */
.hljs-keyword,
.hljs-selector-tag {
color: #8959a8;
}
.hljs {
display: block;
overflow-x: auto;
color: #4d4d4c;
padding: 0.5em;
}
.hljs-emphasis {
font-style: italic;
}
.hljs-strong {
font-weight: bold;
}
</style>
<style>
/*
* Markdown PDF CSS
*/
body {
font-family: -apple-system, BlinkMacSystemFont, "Segoe WPC", "Segoe UI", "Ubuntu", "Droid Sans", sans-serif, "Meiryo";
padding: 0 12px;
}
pre {
background-color: #f8f8f8;
border: 1px solid #cccccc;
border-radius: 3px;
overflow-x: auto;
white-space: pre-wrap;
overflow-wrap: break-word;
}
pre:not(.hljs) {
padding: 23px;
line-height: 19px;
}
blockquote {
background: rgba(127, 127, 127, 0.1);
border-color: rgba(0, 122, 204, 0.5);
}
.emoji {
height: 1.4em;
}
code {
font-size: 14px;
line-height: 19px;
}
/* for inline code */
:not(pre):not(.hljs) > code {
color: #C9AE75; /* Change the old color so it seems less like an error */
font-size: inherit;
}
/* Page Break : use <div class="page"/> to insert page break
-------------------------------------------------------- */
.page {
page-break-after: always;
}
</style>
<script src="https://unpkg.com/mermaid/dist/mermaid.min.js"></script>
</head>
<body>
<script>
mermaid.initialize({
startOnLoad: true,
theme: document.body.classList.contains('vscode-dark') || document.body.classList.contains('vscode-high-contrast')
? 'dark'
: 'default'
});
</script>
<h1 style="background-color:tomato;">Background of the Data</h1>
<ul>
<li>Data Source: https://catalog.data.gov/dataset/consumer-complaint-database</li>
</ul>
<p>The dataset is obtained from the public data from data.gov website under the domain <code>consumer-complaint-database</code>. The Consumer Complaint Database is a collection of complaints about consumer financial products and services that we sent to companies for response. The database generally updates daily. So, each day when we download the dataset, it may be larger than the previous dataset.</p>
<p>The dataset has above 1 million rows and 18 columns out of which, for the text data category classification, we are only interested in two features: <code>Product</code> and <code>Consumer complaint narrative</code>.</p>
<h1 style="background-color:tomato;">Business Problem</h1>
<p>This project aims to accurately classify the Product category of the complaint. There are more than 10 categories of the product such as <code>Mortgage</code>, <code>Debt collection</code> and so on. Our aim is to read the text complaint and classify as on of these category.</p>
<p style="color:green;">NOTE</p>
<p>Originally there are more than 10 categories in original database, but some of the categories are
ambiguous, such as there are three different categories <code>Credit card</code>, <code>Prepaid card</code>, and <code>Credit card or prepaid card</code>. If we have a given complaint about credit card, what should it be classified as? <code>Credit card</code> or <code>Credit card and prepaid card</code> ? To avoid this problem the ambiguous categories are merged into one single categories and finally we have only 10 different categories.
For a sample of 2,000 data, the category distribution looks like this:
<img src="images/labels.png" alt=""></p>
<h1 style="background-color:tomato;">Text Data Cleaning</h1>
<p>Usually the written text is full of informal language and requires cleaning the text before we proceed with
analyzing the text. For example, we need to remove the STOPWORDS and expand the contractions.
Data cleaning strategy:</p>
<ol>
<li>split combined text: <code>areYou</code> ==> <code>are You</code></li>
<li>lowercase: <code>You</code> ==> <code>you</code></li>
<li>expand apostrophes: <code>you're</code> ==> <code>you are</code></li>
<li>remove punctuation: <code>hi !</code> ==> <code>hi</code></li>
<li>remove digits: <code>gr8</code> ==> <code>gr</code></li>
<li>remove repeated substring: <code>ha ha</code> ==> <code>ha</code></li>
<li>remove stop words: <code>I am good</code> ==> <code>good</code></li>
<li>lemmatize: <code>apples</code> ==> <code>apple</code></li>
</ol>
<h1 style="background-color:tomato;">Tf-idf:</h1>
<p>For the text processing tasks (NLP), we usually use a method called <code>Term Frequency - Inverse Document Frequency</code>.</p>
<p style="color:green;">Term Frequency:</p>
<p>This gives how often a given word appears within a document.</p>
<pre class="hljs"><code><div>
TF = Number of times the term appears in the doc
----------------------------------------------
Total number of words in the doc
</div></code></pre>
<p style="color:green;">Inverse Document Frequency:</p>
<p>This gives how often the word appears across the documents.
If a term is very common among documents (e.g.,<code>the</code>, <code>a</code>, <code>is</code>),
then we have low IDF score.</p>
<pre class="hljs"><code><div> Number of docs the term appears
DF = -----------------------------------
Total number docs in the corpus
But, conventionally, document frequency (Df) is defined as log of ratio,
Number of docs the term appears
DF = ln ( ----------------------------------)
Total number docs in the corpus
</div></code></pre>
<p style="color:green;">Term Frequency – Inverse Document Frequency TF-IDF:</p>
<p>TF-IDF is the product of the TF and IDF scores of the term.</p>
<pre class="hljs"><code><div> TF
TF-IDF = ------
DF
</div></code></pre>
<h1 style="background-color:tomato;">Top N correlated terms per category</h1>
<p>We can use scikitlearn text vectorizer class <code>sklearn.feature_extraction.text.TfidfVectorizer</code> to get
the vectorized form of given text data. Then using feature selection (<code>sklearn.feature_selection.chi2</code>) we get
following top most unigrams and bigrams for each categories:
<img src="images/top_correlated_terms.png" alt=""></p>
<h1 style="background-color:tomato;">Modelling Text data</h1>
<p>We can not use the raw text data as the input for <code>scikit-learn</code> classifiers.
We first need to vectorize them and convert the words to number. Here, in this
project I have used the Tf-idf vectorizer with ngram of (1,2) and tried various
classifiers. Among many classifiers, I found svm.LinearSVC gave me the best accuracy.
For the 2019 data with sampling of 2000 samples with random seed of 100, I got the
accuracy of 0.8125 for the test data. For the full data of 2019 (124,907 almost 125k)
after splitting train-test as 80%-20%, I got the accuracy of 0.8068.</p>
<h1 style="background-color:tomato;">Model Evaluation</h1>
<p><img src="images/classification_report.png" alt="">
<img src="images/confusion_matrix.png" alt="">
<img src="images/roc_auc.png" alt="">
<img src="images/precision_recall.png" alt="">
<img src="images/class_prediction_error.png" alt=""></p>
<h1 style="background-color:tomato;">Big Data Analysis</h1>
<p>Here, we have so far used only the small portion of the data (2,000 samples out of million samples) and used <code>scikit-learn</code> models for the text analysis. But, for the real world data, we may need to use all the data for better performances.</p>
<p>For large data, pandas crashes and we need to look for alternative methods such as Amazon AWS or IBM Watson. Also, we can use the open source modules such as <code>dask</code> or <code>pyspark</code> which can scale up to multiple gigabytes of data. For this project, I have used both <code>pyspark</code> and Amazon AWS servers.</p>
<p style="color:green;">NOTE:</p>
<p><code>Pyspark</code> is an immature library. It was borrowed from scala and many functionalities are still need to be implemented. For example, while reading the <code>complaints.csv</code> file, using pandas we can simply use <code>pd.read_csv</code>, however, pyspark is not sophisticated enough to read the csv file automatically when it has multiline. To circumvent these obstacles we can use spark read option with <code>multiLine=True, escape='"'</code>.</p>
<h1 style="background-color:tomato;">Modelling Pipeline</h1>
<p>For text data processing using <code>pyspark</code>, here I have used following pipelines:</p>
<pre class="hljs"><code><div><span class="hljs-keyword">from</span> pyspark.ml.feature <span class="hljs-keyword">import</span> Tokenizer,StopWordsRemover,HashingTF,IDF
tokenizer = Tokenizer().setInputCol(<span class="hljs-string">"complaint"</span>).setOutputCol(<span class="hljs-string">"words"</span>)
remover= StopWordsRemover().setInputCol(<span class="hljs-string">"words"</span>).setOutputCol(<span class="hljs-string">"filtered"</span>).setCaseSensitive(<span class="hljs-literal">False</span>)
hashingTF = HashingTF().setNumFeatures(<span class="hljs-number">1000</span>).setInputCol(<span class="hljs-string">"filtered"</span>).setOutputCol(<span class="hljs-string">"rawFeatures"</span>)
idf = IDF().setInputCol(<span class="hljs-string">"rawFeatures"</span>).setOutputCol(<span class="hljs-string">"features"</span>).setMinDocFreq(<span class="hljs-number">0</span>)
</div></code></pre>
</body>
</html>