A PROFILE-BASED ARCHITECTURE FOR A FLEXIBLE AND PERSONALIZED INFORMATION ACCESS

International audienceInformation relevance is an important issue in information access techniques. A solution for improving this relevance is the personalization or adaptation of the answers provided to users. For this purpose we propose, in this article, an information retrieval and recommendation architecture in which any element is described in detail by a profile. The originality of this architecture is at the level of its generic aspects and the numerous possibilities of interactions between complementary profiles derived from a profile generic model. Thereafter, we propose a flexible use of profiles to evaluate the personal relevance of a user or a users group by combining various matchings between profiles criteria