A divergence from randomness framework of WordNet synsets' distribution for word sense disambiguation

Fragos, K; Skourlas, C

dc.contributor.author	Fragos, K	en
dc.contributor.author	Skourlas, C	en
dc.date.accessioned	2014-03-01T02:43:49Z
dc.date.available	2014-03-01T02:43:49Z
dc.date.issued	2006	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/31518
dc.relation.uri	http://www.scopus.com/inward/record.url?eid=2-s2.0-77954136697&partnerID=40&md5=02f55046366786cdd90a5e40a446783a	en
dc.subject	word sense disambiguation	en
dc.subject	Divergence From Randomness	en
dc.subject.other	Divergence from randomness	en
dc.subject.other	Sample data	en
dc.subject.other	Synsets	en
dc.subject.other	Word Sense Disambiguation	en
dc.subject.other	Wordnet	en
dc.subject.other	Linguistics	en
dc.subject.other	Ontology	en
dc.subject.other	Random processes	en
dc.subject.other	Semantic Web	en
dc.subject.other	Targets	en
dc.subject.other	Natural language processing systems	en
dc.title	A divergence from randomness framework of WordNet synsets' distribution for word sense disambiguation	en
heal.type	conferenceItem	en
heal.publicationDate	2006	en
heal.abstract	We describe and experimentally evaluate a method for word sense disambiguation based on measuring the divergence from the randomness of the WordNet synsets' distribution in the context of a word that is to be disambiguated (target word). Firstly, for each word appearing in the context we collect its related synsets from WordNet using WordNet relations, and creating thus the bag of the related synsets for the context. Secondly, for each one of the senses of the target word we study the distribution of its related synsets in the context bag. Assigning a theoretical random process for these distributions and measuring the divergence from the random process we conclude the correct sense of the target word. The method was evaluated on English lexical sample data from the Senseval-2 word sense disambiguation competition, and exhibited remarkable performance compared to / better than most known WordNet relations based measures for word sense disambiguation. Moreover, the method is general and can conduct the disambiguation task assigning any random process for the distribution of the related synsets and using any measure to quantify the divergence from randomness.	en
heal.journalName	Proceedings of the 3rd International Workshop on Natural Language Understanding and Cognitive Science, NLUCS 2006, in Conjunction with ICEIS 2006	en
dc.identifier.spage	71	en
dc.identifier.epage	80	en