Towards Distributed Information Retrieval in the Semantic Web: Query Reformulation Using the oMAP Framework

This paper introduces a general methodology for performing distributed search in the Semantic Web. We propose to define this task as a three steps process, namely resource selection, query reformulation/ontology alignment and rank aggregation/data fusion. For the second problem, we have implemented oMAP, a formal framework for automatically aligning OWL ontologies. In oMAP, different components are combined for finding suitable mapping candidates (together with their weights), and the set of rules with maximum matching probability is selected. Among these components, traditional terminological-based classifiers, machine learning-based classifiers and a new classifier using the structure and the semantics of the OWL ontologies are proposed. oMAP has been evaluated on international test sets.