fork download
copy 
  1. # your code goes here
  2. !wget http://w...content-available-to-author-only...c.jp/~michi/toku2/book_category.tgz
  3. !tar zxf book_category.tgz
  4.  
  5.  
  6. !ls ./book_category
  7. !ls ./book_category | head
  8. !pip install mecab-python3
  9.  
  10. from sklearn.preprocessing import Normalizer
  11.  
  12.  
  13. topics = [
  14. 'computer_graphics',
  15. 'operating_systems',
  16. 'computer_security',
  17. 'application_service',
  18. 'computer_software',
  19. 'artificial_intelligence',
  20. 'search_engine',
  21. 'information_society',
  22. ]
  23.  
  24.  
  25. import glob
  26. import re
  27.  
  28. import MeCab
  29. import pandas as pd
  30.  
  31. from sklearn.feature_extraction.text import TfidfVectorizer, CountVectorizer
  32. from sklearn.preprocessing import LabelEncoder
  33. from sklearn.model_selection import train_test_split
  34. from sklearn.feature_extraction.text import TfidfTransformer
  35. from sklearn.pipeline import make_pipeline
  36. docs = []
  37. for topic in topics:
  38.     for f in glob.glob(f"./book_category/{topic}/*.txt"):
  39.  
  40.         with open(f, "r") as fin:
  41.             body = "\n".join([line.strip() for line in fin if line.strip()])
  42.         docs.append((topic,body))
  43.  
  44.  
  45. df = pd.DataFrame(
  46.         docs,
  47.         columns=["topic","body"],
  48.         dtype="category"
  49. )
  50.  
  51.  
  52.  
  53. print(df.head())
  54.  
  55.  
  56. df.topic.value_counts()
  57.  
  58.  
  59. tagger = MeCab.Tagger("-Owakati")
  60.  
  61. def parse_to_wakati(text):
  62.     return tagger.parse(text).strip()
  63.  
  64. df = df.assign(body_wakati=df.body.apply(parse_to_wakati))
  65.  
  66. print(df.head())
  67.  
  68. print(df.body_wakati.head())
  69.  
  70.  
  71. le = LabelEncoder()
  72. y = le.fit_transform(df.topic)
  73.  
  74.  
  75. print(le.classes_)
  76. print(le.transform(["computer_graphics"]))
  77. print(le.transform(["operating_systems"]))
  78.  
  79.  
  80.  
  81. X_train, X_test, y_train, y_test = train_test_split(
  82.                                       df.body_wakati,  
  83.                                       y, 
  84.                                       test_size=0.2, 
  85.                                       random_state=10, 
  86.                                       shuffle=True
  87. )
  88.  
  89.  
  90. count_vect = CountVectorizer()
  91. X_train_counts = count_vect.fit_transform(df.body_wakati)
  92. tf_transformer = TfidfTransformer(use_idf=False).fit(X_train_counts)
  93. X_train_tf = tf_transformer.transform(X_train_counts)
  94. print(X_train_tf)
  95. print(X_train_tf.shape)
  96.  
  97.  
  98. from sklearn.base import BaseEstimator, TransformerMixin
  99. from sklearn.metrics import classification_report, confusion_matrix
  100. class RulebasedEstimator(BaseEstimator, TransformerMixin):
  101.     def __init__(self, label_encoder):
  102.         self.le = label_encoder
  103.  
  104.     def fit(self, X, y):
  105.         return self
  106.  
  107.     def predict(self, X):
  108.         """ルールを記述"""
  109.         result = []
  110.         for text in X:
  111.             pred = 0
  112.             if re.search(r"(コンピュータ|グラフィックス|computer|graphics)", text):
  113.                 pred = self.le.transform(["computer_graphics"])[0]
  114.             elif re.search(r"(オペレーティング|システム|operating|system)", text):
  115.                 pred = self.le.transform(["operating_systems"])[0]
  116.             elif re.search(r"(暗号|セキュリティ|computer|security)", text):
  117.                 pred = self.le.transform(["computer_security"])[0]
  118.             elif re.search(r"(アプリケーション|サービス)", text):
  119.                 pred = self.le.transform(["application_service"])[0]
  120.             elif re.search(r"(コンピュータ|ソフトウェア|computer|software)", text):
  121.                 pred = self.le.transform(["computer_software"])[0]
  122.             elif re.search(r"(人工|知能)", text):
  123.                 pred = self.le.transform(["artificial_intelligence"])[0]
  124.             elif re.search(r"(検索|エンジン)", text):
  125.                 pred = self.le.transform(["search_engine"])[0]
  126.             elif re.search(r"(情報|社会)", text):
  127.                 pred = self.le.transform(["information_society"])[0]
  128.             result.append(pred)
  129.         return result
  130.  
  131.  
  132. rulebased = RulebasedEstimator(label_encoder=le)
  133. rulebased_pred = rulebased.predict(X_test)
  134.  
  135.  
  136. from sklearn.metrics import confusion_matrix
  137. confusion_matrix(y_test,rulebased_pred)
  138.  
  139. print(classification_report(y_test, rulebased_pred, target_names=le.classes_))
  140.  
  141. from sklearn.decomposition import PCA
  142. from sklearn.preprocessing import StandardScaler
  143.  
  144. from sklearn.linear_model import SGDClassifier
  145. #pca = make_pipeline(StandardScaler(),PCA(n_components=2, random_state=0))
  146.  
  147. from sklearn.decomposition import TruncatedSVD
  148.  
  149. svd = TruncatedSVD(n_components=1693 , n_iter=7, random_state=42)
  150.  
  151.  
  152. text_clf = Pipeline([
  153.     ('tfidf', TfidfVectorizer()),
  154.     ('std',StandardScaler()),
  155.     ('PCA', TruncatedSVD(n_components=1693 , n_iter=7, random_state=42)),
  156.     ('clf', SGDClassifier(loss='hinge', penalty='l2',
  157.                           alpha=1e-3, random_state=90,
  158.                           max_iter=100, tol=None)),
  159. ])
  160.  
  161. text_clf.fit(X_train,y_train)
  162.  
  163. predicted = text_clf.predict(X_test)
  164.  
  165.  
  166. confusion_matrix(y_test,predicted)
  167.  
  168. print(classification_report(y_test, predicted, target_names=le.classes_))
Compilation error #stdin compilation error #stdout 0s 0KB
comments (?)
stdin
copy 
Standard input is empty
compilation info 
Traceback (most recent call last):
  File "/usr/lib/python3.7/py_compile.py", line 143, in compile
    _optimize=optimize)
  File "<frozen importlib._bootstrap_external>", line 791, in source_to_code
  File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
  File "./prog.py", line 2
    !wget http://www.cs.gunma-u.ac.jp/~michi/toku2/book_category.tgz
    ^
SyntaxError: invalid syntax

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/usr/lib/python3.7/py_compile.py", line 147, in compile
    raise py_exc
py_compile.PyCompileError:   File "./prog.py", line 2
    !wget http://www.cs.gunma-u.ac.jp/~michi/toku2/book_category.tgz
    ^
SyntaxError: invalid syntax
stdout
copy 
Standard output is empty
https://ideone.com/QDfkb4
language:
Python 3 (python  3.12)
created:
6 years ago
visibility:
public

Share or Embed source code

Discover > Sphere Engine API

The brand new service which powers Ideone!

Discover > IDE Widget

Widget for compiling and running the source code in a web browser!