Logical and Conceptual Models for the Integration of Information Retrieval and Database Systems

We present two new approaches to the problem of integrating information retrieval (IR) and database (DB) systems. On the logical level, IR is based on uncertain inference, which is a generalization to the certain inference process employed in DB systems. As an implementation of this concept, we present a probabilistic relational algebra. On the conceptual level, we distinguish between the logical, layout and content structure of DB objects. In the past, DB research used to focus on the logical structure of objects, whereas IR research dealt with the content of documents. By combining these two approaches and incorporating also the layout structure of objects, a conceptual model for integrated IR and DB systems is outlined.