Learning rules for large vocabulary word sense disambiguation

  • Authors:
  • Georgios Paliouras;Vangelis Karkaletsis;Constantine D. Spyropoulos

  • Affiliations:
  • Institute of Informatics & Telecommunications, NCSR "Demokritos", Athens, Greece;Institute of Informatics & Telecommunications, NCSR "Demokritos", Athens, Greece;Institute of Informatics & Telecommunications, NCSR "Demokritos", Athens, Greece

  • Venue:
  • IJCAI'99 Proceedings of the 16th international joint conference on Artificial intelligence - Volume 2
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

Word Sense Disambiguation (WSD) is the process of distinguishing between different senses of a word. In general, the disambiguation rules differ for different words. For this reason, the automatic construction of disambiguation rules is highly desirable. One way to achieve this aim is by applying machine learning techniques to training data containing the various senses of the ambiguous words. In the work presented here, the decision tree learning algorithm C4.5 is applied on a corpus of financial news articles. Instead of concentrating on a small set of ambiguous words, as done in most of the related previous work, all content words of the examined corpus are disambiguated. Furthermore, the effectiveness of word sense disambiguation for different parts of speech (nouns and verbs) is examined empirically.