DMCA
Combining disambiguation techniques to enrich an ontology (2002)
| Venue: | In Proceedings of the Machine Learning and Natural Language Processing for Ontology Engineering Workshop |
| Citations: | 12 - 0 self |
BibTeX
@INPROCEEDINGS{Stevenson02combiningdisambiguation,
author = {Mark Stevenson},
title = {Combining disambiguation techniques to enrich an ontology},
booktitle = {In Proceedings of the Machine Learning and Natural Language Processing for Ontology Engineering Workshop},
year = {2002}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Abstract. This paper presents a method for adding domain information to ontologies using information which is readily available in thesauruses. The approach uses a combination of three existing disambiguation methods from the NLP literature. The approach is implemented on a portion of the WordNet hierarchy with thesaural categories derived from the Cambridge International Dictionary of English. Evaluation is carried out by comparing the system output against a goldstandard derived from SEMCOR, an existing disambiguated corpus, and found to perform well when compared with human judgments. 1 The Tennis Problem WordNet [5] is a broad coverage thesaurus which is now widely used in Natural Language Processing and Information Retrieval. It consists of sets of lexical items (words and
Keyphrases
disambiguation technique domain information thesaural category system output disambiguation method tennis problem wordnet wordnet hierarchy information retrieval cambridge international dictionary broad coverage thesaurus disambiguated corpus natural language processing nlp literature lexical item human judgment
