Identifying concept attributes using a classifier

  • Authors:
  • Massimo Poesio;Abdulrahman Almuhareb

  • Affiliations:
  • University of Essex;University of Essex

  • Venue:
  • DeepLA '05 Proceedings of the ACL-SIGLEX Workshop on Deep Lexical Acquisition
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

We developed a novel classification of concept attributes and two supervised classifiers using this classification to identify concept attributes from candidate attributes extracted from the Web. Our binary (attribute / non-attribute) classifier achieves an accuracy of 81.82% whereas our 5-way classifier achieves 80.35%.