Simple window selection strategies for the simplified lesk algorithm for word sense disambiguation

Francisco Viveros-Jiménez, Alexander Gelbukh, Grigori Sidorov

Producción científica: Capítulo del libro/informe/acta de congresoContribución a la conferenciarevisión exhaustiva

6 Citas (Scopus)

Resumen

The Simplified Lesk Algorithm (SLA) is frequently used for word sense disambiguation. It disambiguates by calculating the overlap of a set of dictionary definitions (senses) and the context words. The algorithm is simple and fast, but it has relatively low accuracy. We propose simple strategies for the context window selection that improve the performance of the SLA: (1) constructing the window only with words that have an overlap with some sense of the target word, (2) excluding the target word itself from matching, and (3) avoiding repetitions in the context window. This paper describes the corresponding experiments. Comparison with other more complex knowledge-based algorithms is presented.

Idioma originalInglés
Título de la publicación alojadaAdvances in Artificial Intelligence and Its Applications - 12th Mexican International Conference on Artificial Intelligence, MICAI 2013, Proceedings
Páginas217-227
Número de páginas11
EdiciónPART 1
DOI
EstadoPublicada - 2013
Evento12th Mexican International Conference on Artificial Intelligence, MICAI 2013 - Mexico City, México
Duración: 24 nov. 201330 nov. 2013

Serie de la publicación

NombreLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NúmeroPART 1
Volumen8265 LNAI
ISSN (versión impresa)0302-9743
ISSN (versión digital)1611-3349

Conferencia

Conferencia12th Mexican International Conference on Artificial Intelligence, MICAI 2013
País/TerritorioMéxico
CiudadMexico City
Período24/11/1330/11/13

Huella

Profundice en los temas de investigación de 'Simple window selection strategies for the simplified lesk algorithm for word sense disambiguation'. En conjunto forman una huella única.

Citar esto