Preferences for Queries in a Mediator Approach

The problem of integrating relevant information obtained from multiple heterogeneous sources is a complex task, with which biologists are now faced. In this paper, we address the problem of querying biomedical databases in a mediator context. We propose to exploit the metadata of the sources to take into account user preferences. The mediator system we present is designed within a tractable logical framework. It allows both transparent and cooperative querying and makes it possible to keep track of the origins of the instances provided as answers. Our proposal is generic in that it is relevant not only for bioinformatics, but could also be applied to other domains for which metadata are available.

• Type: Conference
• Year: 2004
• Where: ECAI
• Authors: Alain Bidault, Sarah Cohen Boulakia, Christine Froidevaux
• Keywords: Account User Preferences, Artificial Intelligence, ECAI 2004, Multiple Heterogeneous Sources, Tractable Logical Framework