Using value-added document representations in INEX

Viewing Information Retrieval from the cognitive viewpoint we generated different functional and cognitive representations of the INEX corpus using both the XML structure and external sources. This included the use of a citation index and intellectually assigned descriptors, and an expansion of the document representation through a domain thesaurus. The aim was to investigate the possible benefits from applying the principle of polyrepresentation [3]. Results showed that neither the descriptors nor the expanded document representation through the thesaurus could improve
results with the natural language queries used. The citation index achieved a similar performance to that obtained using various kinds of titles extracted from the XML structure.

Zitieren

Zitierform:
Zitierform konnte nicht geladen werden.

Rechte

Nutzung und Vervielfältigung:
Alle Rechte vorbehalten