Can ontologies improve Web search engine effectiveness before the advent of the semantic Web?

Jacques RobinFranklin Ramalho

The advent of the semantic web will not be immediate, due to a shortage of knowledge engineering manpower to annotate a critical mass of web documents with concepts from ontologies. In this paper, we argue that widecoverage linguistic ontologies have an alternative usage to improve web search effectiveness while the web remains predominantly textual. They can be used by search engines to perform simple semantic processing, such as automatically expanding the query keywords with alternate wordings of the same or immediately neighboring concepts. We present a conclusive empirical experiment that shows that sizable relative effectiveness gains can be achieved by various expansion strategies for both unbounded and bounded retrieval result sets. This experiment also shows that the effectiveness of such ontology-boosted search on a textual web with no semantic annotations remains far away from the very high precision and recall promise of the semantic web.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: