A Mapping Between Classifiers and Training Conditions for WSD

This paper studies performance of various classifiers for Word Sense Disambiguation considering different training conditions. Our preliminary results indicate that the number and distribution of training examples has a great impact on the resulting precision. The Naïve Bayes method emerged as the most adequate classifier for disambiguating words having few examples.

• Type: Conference
• Year: 2005
• Where: CICLING
• Authors: Aarón Pancardo-Rodríguez, Manuel Montes-y-Gómez, Luis Villaseñor Pineda, Paolo Rosso
• Keywords: CICLING 2005, Naïve Bayes Method, Natural Language Processing, Paper Studies Performance, Word Sense Disambiguation