Abstract de la publi numéro 11091

It is well known that the fundamental intellectual problems of information access are the production and consumption of information. In this paper, we investigate the use of social network of information producers (authors) within relations in data (co-authorship and citation) in order to improve the relevance of information access. Relevance is derived from the network by levraging the usual topical similarity between the query and the document with the target author’s authority. We explore various social network based measures for computing social information importance and show how this kind of contextual information can be incorporated within an information access model. We experiment with a collection issued from SIGIR proceedings and show that combining topical, author and citation based evidences can significantly improve retrieval access precision, measured in terms of mean reciprocal rank.