Using Statistical Term Similarity for Sense Disambiguationin Cross-Language Information Retrieval

  • Authors:
  • Mirna Adriani

  • Affiliations:
  • Department of Computing Science, University of Glasgow, Glasgow G12 8QQ, Scotland. mirna@dcs.gla.ac.uk

  • Venue:
  • Information Retrieval
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

With the increasing availability of machine-readablebilingual dictionaries, dictionary-based automatic query translationhas become a viable approach to Cross-Language Information Retrieval(CLIR). In this approach, resolving term ambiguity is a crucial step.We propose a sense disambiguation technique based on aterm-similarity measure for selecting the right translation sense ofa query term. In addition, we apply a query expansion technique whichis also based on the term similarity measure to improve theeffectiveness of the translation queries. The results of ourIndonesian to English and English to Indonesian CLIR experimentsdemonstrate the effectiveness of the sense disambiguation technique.As for the query expansion technique, it is shown to be effective aslong as the term ambiguity in the queries has been resolved. In theeffort to solve the term ambiguity problem, we discovered thatdifferences in the pattern of word-formation between the twolanguages render query translations from one language to the otherdifficult.