Tag semantics for the retrieval of XML documents

Word Sense Disambiguation (WSD), in the field of Natural Language Processing (NLP), consists in assigning the correct sense (semantics) to a word form (lexeme) by means of the context in which the lexeme is found. In this paper we investigate the possibility of applying WSD techniques to the field of Information Retrieval, especially to the retrieval of XML documents. We consider two methods to automatically assign semantic values to XML tags on the grounds of the tagged text contained. Such methods rely on the bayesian supervised approach and on an automatic unsupervised approach and exploit the WordNet ontology. Results show that the applicability of both methods is hampered by the habit of use abbreviations or shortcuts as tags.

• Type: Conference
• Year: 2003
• Where: ISICT
• Authors: Davide Buscaldi, Giovanna Guerrini, Marco Mesiti, Paolo Rosso
• Keywords: Information Technology, ISICT 2003, Natural Language Processing, Word Sense Disambiguation, WSD Techniques