/
classifier.html
504 lines (466 loc) · 23.4 KB
/
classifier.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
<!doctype html>
<html lang="en">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, minimum-scale=1" />
<meta name="generator" content="pdoc 0.10.0" />
<title>ktrain.text.shallownlp.classifier API documentation</title>
<meta name="description" content="" />
<link rel="preload stylesheet" as="style" href="https://cdnjs.cloudflare.com/ajax/libs/10up-sanitize.css/11.0.1/sanitize.min.css" integrity="sha256-PK9q560IAAa6WVRRh76LtCaI8pjTJ2z11v0miyNNjrs=" crossorigin>
<link rel="preload stylesheet" as="style" href="https://cdnjs.cloudflare.com/ajax/libs/10up-sanitize.css/11.0.1/typography.min.css" integrity="sha256-7l/o7C8jubJiy74VsKTidCy1yBkRtiUGbVkYBylBqUg=" crossorigin>
<link rel="stylesheet preload" as="style" href="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/10.1.1/styles/github.min.css" crossorigin>
<style>:root{--highlight-color:#fe9}.flex{display:flex !important}body{line-height:1.5em}#content{padding:20px}#sidebar{padding:30px;overflow:hidden}#sidebar > *:last-child{margin-bottom:2cm}.http-server-breadcrumbs{font-size:130%;margin:0 0 15px 0}#footer{font-size:.75em;padding:5px 30px;border-top:1px solid #ddd;text-align:right}#footer p{margin:0 0 0 1em;display:inline-block}#footer p:last-child{margin-right:30px}h1,h2,h3,h4,h5{font-weight:300}h1{font-size:2.5em;line-height:1.1em}h2{font-size:1.75em;margin:1em 0 .50em 0}h3{font-size:1.4em;margin:25px 0 10px 0}h4{margin:0;font-size:105%}h1:target,h2:target,h3:target,h4:target,h5:target,h6:target{background:var(--highlight-color);padding:.2em 0}a{color:#058;text-decoration:none;transition:color .3s ease-in-out}a:hover{color:#e82}.title code{font-weight:bold}h2[id^="header-"]{margin-top:2em}.ident{color:#900}pre code{background:#f8f8f8;font-size:.8em;line-height:1.4em}code{background:#f2f2f1;padding:1px 4px;overflow-wrap:break-word}h1 code{background:transparent}pre{background:#f8f8f8;border:0;border-top:1px solid #ccc;border-bottom:1px solid #ccc;margin:1em 0;padding:1ex}#http-server-module-list{display:flex;flex-flow:column}#http-server-module-list div{display:flex}#http-server-module-list dt{min-width:10%}#http-server-module-list p{margin-top:0}.toc ul,#index{list-style-type:none;margin:0;padding:0}#index code{background:transparent}#index h3{border-bottom:1px solid #ddd}#index ul{padding:0}#index h4{margin-top:.6em;font-weight:bold}@media (min-width:200ex){#index .two-column{column-count:2}}@media (min-width:300ex){#index .two-column{column-count:3}}dl{margin-bottom:2em}dl dl:last-child{margin-bottom:4em}dd{margin:0 0 1em 3em}#header-classes + dl > dd{margin-bottom:3em}dd dd{margin-left:2em}dd p{margin:10px 0}.name{background:#eee;font-weight:bold;font-size:.85em;padding:5px 10px;display:inline-block;min-width:40%}.name:hover{background:#e0e0e0}dt:target .name{background:var(--highlight-color)}.name > span:first-child{white-space:nowrap}.name.class > span:nth-child(2){margin-left:.4em}.inherited{color:#999;border-left:5px solid #eee;padding-left:1em}.inheritance em{font-style:normal;font-weight:bold}.desc h2{font-weight:400;font-size:1.25em}.desc h3{font-size:1em}.desc dt code{background:inherit}.source summary,.git-link-div{color:#666;text-align:right;font-weight:400;font-size:.8em;text-transform:uppercase}.source summary > *{white-space:nowrap;cursor:pointer}.git-link{color:inherit;margin-left:1em}.source pre{max-height:500px;overflow:auto;margin:0}.source pre code{font-size:12px;overflow:visible}.hlist{list-style:none}.hlist li{display:inline}.hlist li:after{content:',\2002'}.hlist li:last-child:after{content:none}.hlist .hlist{display:inline;padding-left:1em}img{max-width:100%}td{padding:0 .5em}.admonition{padding:.1em .5em;margin-bottom:1em}.admonition-title{font-weight:bold}.admonition.note,.admonition.info,.admonition.important{background:#aef}.admonition.todo,.admonition.versionadded,.admonition.tip,.admonition.hint{background:#dfd}.admonition.warning,.admonition.versionchanged,.admonition.deprecated{background:#fd4}.admonition.error,.admonition.danger,.admonition.caution{background:lightpink}</style>
<style media="screen and (min-width: 700px)">@media screen and (min-width:700px){#sidebar{width:30%;height:100vh;overflow:auto;position:sticky;top:0}#content{width:70%;max-width:100ch;padding:3em 4em;border-left:1px solid #ddd}pre code{font-size:1em}.item .name{font-size:1em}main{display:flex;flex-direction:row-reverse;justify-content:flex-end}.toc ul ul,#index ul{padding-left:1.5em}.toc > ul > li{margin-top:.5em}}</style>
<style media="print">@media print{#sidebar h1{page-break-before:always}.source{display:none}}@media print{*{background:transparent !important;color:#000 !important;box-shadow:none !important;text-shadow:none !important}a[href]:after{content:" (" attr(href) ")";font-size:90%}a[href][title]:after{content:none}abbr[title]:after{content:" (" attr(title) ")"}.ir a:after,a[href^="javascript:"]:after,a[href^="#"]:after{content:""}pre,blockquote{border:1px solid #999;page-break-inside:avoid}thead{display:table-header-group}tr,img{page-break-inside:avoid}img{max-width:100% !important}@page{margin:0.5cm}p,h2,h3{orphans:3;widows:3}h1,h2,h3,h4,h5,h6{page-break-after:avoid}}</style>
<script defer src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/10.1.1/highlight.min.js" integrity="sha256-Uv3H6lx7dJmRfRvH8TH6kJD1TSK1aFcwgx+mdg3epi8=" crossorigin></script>
<script>window.addEventListener('DOMContentLoaded', () => hljs.initHighlighting())</script>
</head>
<body>
<main>
<article id="content">
<header>
<h1 class="title">Module <code>ktrain.text.shallownlp.classifier</code></h1>
</header>
<section id="section-intro">
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">from . import utils as U
from .imports import *
__all__ = ["NBSVM"]
class Classifier:
def __init__(self, model=None):
"""
instantiate a classifier with an optional previously-saved model
"""
self.model = None
def create_model(self, ctype, texts, hp_dict={}, ngram_range=(1, 3), binary=True):
"""
```
create a model
Args:
ctype(str): one of {'nbsvm', 'logreg', 'sgdclassifier'}
texts(list): list of texts
hp_dict(dict): dictionary of hyperparameters to use for the ctype selected.
hp_dict can also be used to supply arguments to CountVectorizer
ngram_range(tuple): default ngram_range.
overridden if 'ngram_range' in hp_dict
binary(bool): default value for binary argument to CountVectorizer.
overridden if 'binary' key in hp_dict
```
"""
lang = U.detect_lang(texts)
if U.is_chinese(lang):
token_pattern = r"(?u)\b\w+\b"
else:
token_pattern = r"\w+|[%s]" % string.punctuation
if ctype == "nbsvm":
clf = NBSVM(
C=hp_dict.get("C", 0.01),
alpha=hp_dict.get("alpha", 0.75),
beta=hp_dict.get("beta", 0.25),
fit_intercept=hp_dict.get("fit_intercept", False),
)
elif ctype == "logreg":
clf = LogisticRegression(
C=hp_dict.get("C", 0.1),
dual=hp_dict.get("dual", True),
penalty=hp_dict.get("penalty", "l2"),
tol=hp_dict.get("tol", 1e-4),
intercept_scaling=hp_dict.get("intercept_scaling", 1),
solver=hp_dict.get("solver", "liblinear"),
max_iter=hp_dict.get("max_iter", 100),
multi_class=hp_dict.get("multi_class", "auto"),
warm_start=hp_dict.get("warm_start", False),
n_jobs=hp_dict.get("n_jobs", None),
l1_ratio=hp_dict.get("l1_ratio", None),
random_state=hp_dict.get("random_state", 42),
class_weight=hp_dict.get("class_weight", None),
)
elif ctype == "sgdclassifier":
clf = SGDClassifier(
loss=hp_dict.get("loss", "hinge"),
penalty=hp_dict.get("penalty", "l2"),
alpha=hp_dict.get("alpha", 1e-3),
random_state=hp_dict.get("random_state", 42),
max_iter=hp_dict.get("max_iter", 5), # scikit-learn default is 1000
tol=hp_dict.get("tol", None),
l1_ratio=hp_dict.get("l1_ratio", 0.15),
fit_intercept=hp_dict.get("fit_intercept", True),
episilon=hp_dict.get("epsilon", 0.1),
n_jobs=hp_dict.get("n_jobs", None),
learning_rate=hp_dict.get("learning_rate", "optimal"),
eta0=hp_dict.get("eta0", 0.0),
power_t=hp_dict.get("power_t", 0.5),
early_stopping=hp_dict.get("early_stopping", False),
validation_fraction=hp_dict.get("validation_fraction", 0.1),
n_iter_no_change=hp_dict.get("n_iter_no_change", 5),
warm_start=hp_dict.get("warm_start", False),
average=hp_dict.get("average", False),
class_weight=hp_dict.get("class_weight", None),
)
else:
raise ValueError("Unknown ctype: %s" % (ctype))
self.model = Pipeline(
[
(
"vect",
CountVectorizer(
ngram_range=hp_dict.get("ngram_range", ngram_range),
binary=hp_dict.get("binary", binary),
token_pattern=token_pattern,
max_features=hp_dict.get("max_features", None),
max_df=hp_dict.get("max_df", 1.0),
min_df=hp_dict.get("min_df", 1),
stop_words=hp_dict.get("stop_words", None),
lowercase=hp_dict.get("lowercase", True),
strip_accents=hp_dict.get("strip_accents", None),
encoding=hp_dict.get("encoding", "utf-8"),
),
),
("clf", clf),
]
)
return
@classmethod
def load_texts_from_folder(
cls, folder_path, subfolders=None, shuffle=True, encoding=None
):
"""
```
load text files from folder
Args:
folder_path(str): path to folder containing documents
The supplied folder should contain a subfolder
for each category, which will be used as the class label
subfolders(list): list of subfolders under folder_path to consider
Example: If folder_path contains subfolders pos, neg, and
unlabeled, then unlabeled folder can be ignored by
setting subfolders=['pos', 'neg']
shuffle(bool): If True, list of texts will be shuffled
encoding(str): encoding to use. default:None (auto-detected)
Returns:
tuple: (texts, labels, label_names)
```
"""
bunch = load_files(folder_path, categories=subfolders, shuffle=shuffle)
texts = bunch.data
labels = bunch.target
label_names = bunch.target_names
# print('target names:')
# for idx, label_name in enumerate(bunch.target_names):
# print('\t%s:%s' % (idx, label_name))
# decode based on supplied encoding
if encoding is None:
encoding = U.detect_encoding(texts)
if encoding != "utf-8":
print("detected encoding: %s" % (encoding))
try:
texts = [text.decode(encoding) for text in texts]
except:
print(
"Decoding with %s failed 1st attempt - using %s with skips"
% (encoding, encoding)
)
texts = U.decode_by_line(texts, encoding=encoding)
return (texts, labels, label_names)
@classmethod
def load_texts_from_csv(
cls,
csv_filepath,
text_column="text",
label_column="label",
sep=",",
encoding=None,
):
"""
```
load text files from csv file
CSV should have at least two columns.
Example:
Text | Label
I love this movie. | positive
I hated this movie.| negative
Args:
csv_filepath(str): path to CSV file
text_column(str): name of column containing the texts. default:'text'
label_column(str): name of column containing the labels in string format
default:'label'
sep(str): character that separates columns in CSV. default:','
encoding(str): encoding to use. default:None (auto-detected)
Returns:
tuple: (texts, labels, label_names)
```
"""
if encoding is None:
with open(csv_filepath, "rb") as f:
encoding = U.detect_encoding([f.read()])
if encoding != "utf-8":
print("detected encoding: %s (if wrong, set manually)" % (encoding))
import pandas as pd
df = pd.read_csv(csv_filepath, encoding=encoding, sep=sep)
texts = df[text_column].fillna("fillna").values
labels = df[label_column].values
le = LabelEncoder()
le.fit(labels)
labels = le.transform(labels)
return (texts, labels, le.classes_)
def fit(self, x_train, y_train, ctype="logreg"):
"""
```
train a classifier
Args:
x_train(list or np.ndarray): training texts
y_train(np.ndarray): training labels
ctype(str): One of {'logreg', 'nbsvm', 'sgdclassifier'}. default:nbsvm
```
"""
lang = U.detect_lang(x_train)
if U.is_chinese(lang):
x_train = U.split_chinese(x_train)
if self.model is None:
self.create_model(ctype, x_train)
self.model.fit(x_train, y_train)
return self
def predict(self, x_test, return_proba=False):
"""
```
make predictions on text data
Args:
x_test(list or np.ndarray or str): array of texts on which to make predictions or a string representing text
```
"""
if return_proba and not hasattr(self.model["clf"], "predict_proba"):
raise ValueError(
"%s does not support predict_proba" % (type(self.model["clf"]).__name__)
)
if isinstance(x_test, str):
x_test = [x_test]
lang = U.detect_lang(x_test)
if U.is_chinese(lang):
x_test = U.split_chinese(x_test)
if self.model is None:
raise ValueError("model is None - call fit or load to set the model")
if return_proba:
predicted = self.model.predict_proba(x_test)
else:
predicted = self.model.predict(x_test)
if len(predicted) == 1:
predicted = predicted[0]
return predicted
def predict_proba(self, x_test):
"""
predict_proba
"""
return self.predict(x_test, return_proba=True)
def evaluate(self, x_test, y_test):
"""
```
evaluate
Args:
x_test(list or np.ndarray): training texts
y_test(np.ndarray): training labels
```
"""
predicted = self.predict(x_test)
return np.mean(predicted == y_test)
def save(self, filename):
"""
save model
"""
dump(self.model, filename)
def load(self, filename):
"""
load model
"""
self.model = load(filename)
def grid_search(self, params, x_train, y_train, n_jobs=-1):
"""
```
Performs grid search to find optimal set of hyperparameters
Args:
params (dict): A dictionary defining the space of the search.
Example for finding optimal value of alpha in NBSVM:
parameters = {
#'clf__C': (1e0, 1e-1, 1e-2),
'clf__alpha': (0.1, 0.2, 0.4, 0.5, 0.75, 0.9, 1.0),
#'clf__fit_intercept': (True, False),
#'clf__beta' : (0.1, 0.25, 0.5, 0.9)
}
n_jobs(int): number of jobs to run in parallel. default:-1 (use all processors)
```
"""
gs_clf = GridSearchCV(self.model, params, n_jobs=n_jobs)
gs_clf = gs_clf.fit(x_train, y_train)
# gs_clf.best_score_
for param_name in sorted(params.keys()):
print("%s: %r" % (param_name, gs_clf.best_params_[param_name]))
return
class NBSVM(BaseEstimator, LinearClassifierMixin, SparseCoefMixin):
def __init__(self, alpha=1, C=1, beta=0.25, fit_intercept=False):
self.alpha = alpha
self.C = C
self.beta = beta
self.fit_intercept = fit_intercept
def fit(self, X, y):
self.classes_ = np.unique(y)
if len(self.classes_) == 2:
coef_, intercept_ = self._fit_binary(X, y)
self.coef_ = coef_
self.intercept_ = intercept_
else:
coef_, intercept_ = zip(
*[self._fit_binary(X, y == class_) for class_ in self.classes_]
)
self.coef_ = np.concatenate(coef_)
self.intercept_ = np.array(intercept_).flatten()
return self
def _fit_binary(self, X, y):
p = np.asarray(self.alpha + X[y == 1].sum(axis=0)).flatten()
q = np.asarray(self.alpha + X[y == 0].sum(axis=0)).flatten()
r = np.log(p / np.abs(p).sum()) - np.log(q / np.abs(q).sum())
b = np.log((y == 1).sum()) - np.log((y == 0).sum())
if isinstance(X, spmatrix):
indices = np.arange(len(r))
r_sparse = coo_matrix((r, (indices, indices)), shape=(len(r), len(r)))
X_scaled = X * r_sparse
else:
X_scaled = X * r
lsvc = LinearSVC(
C=self.C, fit_intercept=self.fit_intercept, max_iter=10000
).fit(X_scaled, y)
mean_mag = np.abs(lsvc.coef_).mean()
coef_ = (1 - self.beta) * mean_mag * r + self.beta * (r * lsvc.coef_)
intercept_ = (1 - self.beta) * mean_mag * b + self.beta * lsvc.intercept_
return coef_, intercept_</code></pre>
</details>
</section>
<section>
</section>
<section>
</section>
<section>
</section>
<section>
<h2 class="section-title" id="header-classes">Classes</h2>
<dl>
<dt id="ktrain.text.shallownlp.classifier.NBSVM"><code class="flex name class">
<span>class <span class="ident">NBSVM</span></span>
<span>(</span><span>alpha=1, C=1, beta=0.25, fit_intercept=False)</span>
</code></dt>
<dd>
<div class="desc"><p>Base class for all estimators in scikit-learn.</p>
<h2 id="notes">Notes</h2>
<p>All estimators should specify all the parameters that can be set
at the class level in their <code>__init__</code> as explicit keyword
arguments (no <code>*args</code> or <code>**kwargs</code>).</p></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">class NBSVM(BaseEstimator, LinearClassifierMixin, SparseCoefMixin):
def __init__(self, alpha=1, C=1, beta=0.25, fit_intercept=False):
self.alpha = alpha
self.C = C
self.beta = beta
self.fit_intercept = fit_intercept
def fit(self, X, y):
self.classes_ = np.unique(y)
if len(self.classes_) == 2:
coef_, intercept_ = self._fit_binary(X, y)
self.coef_ = coef_
self.intercept_ = intercept_
else:
coef_, intercept_ = zip(
*[self._fit_binary(X, y == class_) for class_ in self.classes_]
)
self.coef_ = np.concatenate(coef_)
self.intercept_ = np.array(intercept_).flatten()
return self
def _fit_binary(self, X, y):
p = np.asarray(self.alpha + X[y == 1].sum(axis=0)).flatten()
q = np.asarray(self.alpha + X[y == 0].sum(axis=0)).flatten()
r = np.log(p / np.abs(p).sum()) - np.log(q / np.abs(q).sum())
b = np.log((y == 1).sum()) - np.log((y == 0).sum())
if isinstance(X, spmatrix):
indices = np.arange(len(r))
r_sparse = coo_matrix((r, (indices, indices)), shape=(len(r), len(r)))
X_scaled = X * r_sparse
else:
X_scaled = X * r
lsvc = LinearSVC(
C=self.C, fit_intercept=self.fit_intercept, max_iter=10000
).fit(X_scaled, y)
mean_mag = np.abs(lsvc.coef_).mean()
coef_ = (1 - self.beta) * mean_mag * r + self.beta * (r * lsvc.coef_)
intercept_ = (1 - self.beta) * mean_mag * b + self.beta * lsvc.intercept_
return coef_, intercept_</code></pre>
</details>
<h3>Ancestors</h3>
<ul class="hlist">
<li>sklearn.base.BaseEstimator</li>
<li>sklearn.linear_model._base.LinearClassifierMixin</li>
<li>sklearn.base.ClassifierMixin</li>
<li>sklearn.linear_model._base.SparseCoefMixin</li>
</ul>
<h3>Methods</h3>
<dl>
<dt id="ktrain.text.shallownlp.classifier.NBSVM.fit"><code class="name flex">
<span>def <span class="ident">fit</span></span>(<span>self, X, y)</span>
</code></dt>
<dd>
<div class="desc"></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">def fit(self, X, y):
self.classes_ = np.unique(y)
if len(self.classes_) == 2:
coef_, intercept_ = self._fit_binary(X, y)
self.coef_ = coef_
self.intercept_ = intercept_
else:
coef_, intercept_ = zip(
*[self._fit_binary(X, y == class_) for class_ in self.classes_]
)
self.coef_ = np.concatenate(coef_)
self.intercept_ = np.array(intercept_).flatten()
return self</code></pre>
</details>
</dd>
</dl>
</dd>
</dl>
</section>
</article>
<nav id="sidebar">
<h1>Index</h1>
<div class="toc">
<ul></ul>
</div>
<ul id="index">
<li><h3>Super-module</h3>
<ul>
<li><code><a title="ktrain.text.shallownlp" href="index.html">ktrain.text.shallownlp</a></code></li>
</ul>
</li>
<li><h3><a href="#header-classes">Classes</a></h3>
<ul>
<li>
<h4><code><a title="ktrain.text.shallownlp.classifier.NBSVM" href="#ktrain.text.shallownlp.classifier.NBSVM">NBSVM</a></code></h4>
<ul class="">
<li><code><a title="ktrain.text.shallownlp.classifier.NBSVM.fit" href="#ktrain.text.shallownlp.classifier.NBSVM.fit">fit</a></code></li>
</ul>
</li>
</ul>
</li>
</ul>
</nav>
</main>
<footer id="footer">
<p>Generated by <a href="https://pdoc3.github.io/pdoc" title="pdoc: Python API documentation generator"><cite>pdoc</cite> 0.10.0</a>.</p>
</footer>
</body>
</html>