Genres for Web Page Classification

1
Genres for Web Page Classification

• Vedrana Vidulin, Mitja Lutrek, Matja Gams
• Department of Intelligent Systems
• Joef Stefan Institute
• vedrana.vidulin_at_ijs.si

2
Why genres?

• Genre could be described as a style of a web page.

3
Training a Genre Classifier

• Corpus of web pages annotated with genres
• Set of features
• Machine learning algorithm

4
20-Genre Corpus

• 1539 web pages divided into 20 genres
• Multilabeled corpus

5
The Set of Features

• URL features
• https, www, document type, URL words
• HTML features
• number of hyperlinks to the same domain / total
  number of hyperlinks, number of tags / total
  number of tags for e.g. tag group Interaction
• Text features
• average number of characters per word, function
  words, punctuation symbols, content words, number
  of declarative sentences / total number of
  sentences

6
Machine Learning

• Multilabeled classification
• 20 binary machine learning tasks
• Machine learning algorithms
• Decision Tree
• Bagging

7
Results