Modelling Vague Content and Structure Querying in XML Retrieval with a Probabilistic Object-Relational Framework

Many XML retrieval applications require relevance-oriented ranking of retrieved elements in order to capture the vagueness inherent to the information retrieval process. This relevance-oriented ranking should not only support vagueness at the content level, but also at the structural level. In this paper, we use a probabilistic object-relational framework to model representation and retrieval strategies that take into account vagueness at both content and structure level. Our approach makes use of established database technology combined with sound probability theory, thus allowing for fast and flexible prototyping of various representation and retrieval strategies.