Cross-Language Information Retrieval (CLIR) systems extend classic information retrieval mechanisms for allowing users to query across languages, i.e., to retrieve documents written in languages different from the language used for query formulation. In this paper, we present a CLIR system exploiting multilingual ontologies for enriching documents representation with multilingual semantic information during the indexing phase and for mapping query fragments to concepts during the retrieval phase. This system has been applied on a domainspecific document collection and the contribution of the ontologies to the CLIR system has been evaluated in conjunction with the use of both Microsoft Bing and Google Translate translation services. Results demonstrate that the use of domain-specific resources leads to a significant improvement of CLIR system performance.
Using Semantic and Domain-Based Information in CLIR Systems
Dragoni, Mauro;Di Francescomarino, Chiara
2014-01-01
Abstract
Cross-Language Information Retrieval (CLIR) systems extend classic information retrieval mechanisms for allowing users to query across languages, i.e., to retrieve documents written in languages different from the language used for query formulation. In this paper, we present a CLIR system exploiting multilingual ontologies for enriching documents representation with multilingual semantic information during the indexing phase and for mapping query fragments to concepts during the retrieval phase. This system has been applied on a domainspecific document collection and the contribution of the ontologies to the CLIR system has been evaluated in conjunction with the use of both Microsoft Bing and Google Translate translation services. Results demonstrate that the use of domain-specific resources leads to a significant improvement of CLIR system performance.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
