Approximate Information Filtering on the Semantic Web

Abstract

Facing the increasing amount of information available on the World Wide Web, intelligent techniques for content-based information filtering gain more and more importance. Conventional approaches using keyword- or text-based retrieval methods have been developed that perform reasonably well. However, these approaches have problems with ambiguous and imprecise information. The semantic web that aims at supplementing information sources with a formal specification of its meaning using ontologies can potentially help to overcome this problem. At the moment, however, the semantic web still suffers from its own problems in terms of heterogeneous ontologies and the need to relate them to each other. In this paper, we argue that we can overcome this problem by using shared vocabularies, a standardized language for encoding ontology that supports basic terminological reasoning (in this case DAML+OIL) and techniques from approximate reasoning. We introduce the approach on an informal level using didactic example and give a formal characterization of the method that include correctness proofs for the problem of information filtering.