The Semantic Dimension in Information Retrieval, from Document Indexing to Query Reformulation
Abdelkrim Bouramoul · International Conference on Information Systems · 2011
In the context of this research, we present an approach for representing the semantic content of documents and guiding the automatic query reformulation using a domain ontology. The aim is to improve the performance of information retrieval systems. In order to operationalize our proposal, the development of a set of external resources was needed, so we have constructed the AnimOnto' domain ontology relating to the animals domain and a document base that covers the same field. They were used to test and validate our proposal. Specifically, we propose in this paper a general architecture based on three complementary processes, this architecture uses the ontology during the semantic indexing stage and in the query reformulation phase. We also describe the 'AnimeSe Finder' tool (Animal Semantic Finder). The latter has the advantage of being generic and adaptable to other search types. It is possible to use another ontology with another document base for a new domain, to exploit the general functionalities offered by the 'AnimeSe Finder' tool.