Home // SEMAPRO 2010, The Fourth International Conference on Advances in Semantic Processing // View article

Exploiting WordNet glosses to disambiguate nouns through verbs

Authors:
Donato Barbagallo
Leonardo Bruni
Chiara Francalanci

Keywords: word sense disambiguation; WordNet; genetic algorithm

Abstract:
This paper presents an unsupervised graph-based algorithm for word sense disambiguation based on WordNet glosses. The algorithm exploits the contributions of verbs in identifying the correct senses of nouns. Due to the complexity of WordNet’s semantic network, we have defined disambiguation as a similarity optimization problem and solved it through a genetic algorithm. Testing compares the performance of our algorithm with that of a traditional method based on Wu-Palmer similarity measure. Our approach shows an overall precision of about 68% and a statistically significant average increase of precision of about 3% with respect to the traditional algorithm.

Pages: 173 to 178

Copyright: Copyright (c) IARIA, 2010

Publication date: October 25, 2010

Published in: conference

ISSN: 2308-4510

ISBN: 978-1-61208-104-5

Location: Florence, Italy

Dates: from October 25, 2010 to October 30, 2010