Automatic Noun Sense Disambiguation

This paper explores a fully automatic knowledge-based method which performs the noun sense disambiguation relying only on the WordNet ontology. The basis of the method is the idea of conceptual density, that is, the correlation between the sense of a given word and its context. A new formula for calculating the conceptual density was proposed and was evaluated on the SemCor corpus. 1 An Extension of the Conceptual Density The task of Word Sense Disambiguation (WSD) consists of examining word tokens and specifying exactly which sense of each word is being used. The WordNet (WN) ontology, based on synsets (sets of synonyms), is the external lexical resource which is often used to perform the WSD task. In most of the WSD approaches, a word is disambiguated along with a portion of the text in which it is embedded, that is, its context. When the initial input source of information (i.e., the word and its context) is processed only together with the lexical knowledge source (e.g. WN), a fully automatic method which does not require any kind of training process is needed to perform WSD. Conceptual Density (CD) is a measure of the correlation among the sense of a given word and its context. The foundation of this measure is the Conceptual Distance, defined as the length of the shortest path which connects two concepts in a hierarchical semantic net. The starting point for our work was the CD formula of Agirre and Rigau [1], which compares areas of subhierarchies: CD(c,m) = ∑m−1 i=0 nhyp i ∑h−1 i=0 nhyp i (1) where c is the synset at the top of subhierarchy, m the number of word senses falling within a subhierarchy, h the height of the subhierarchy, and nhyp the This work was supported by the Spanish Research Projects CICYT TIC2000-0664C02 and TIC2000-1599-C01-01, and the VIDI of the Univ. Politécnica Valencia. A. Gelbukh (Ed.): CICLing 2003, LNCS 2588, pp. 273–276, 2003. c © Springer-Verlag Berlin Heidelberg 2003