Ontology mapping in community support systems

Communities have been shown to be a very valuable source of knowledge. Thus, community support has become an important issue for Knowledge Management efforts. Community support systems must reflect the loose coupling of community members as well as the heterogeneity of communities. The heterogeneity of community members brings with it a semantic heterogeneity of terminology and perspective on the available knowledge in the community. A solution for this problem lies in the use of a centralized community ontology on which community support services can be based for semantic interoperability. We introduce an agent-based architecture for community Knowledge Management support. Exemplary support services are presented and it is shown how those services can be implemented, based on a centralized ontology. From a usability perspective, however, an approach with a centralized ontology has a number of disadvantages. We will discuss these disadvantages and present a concept for community support with distributed ontologies. The distributed ontologies paradigm is shown to be superior in terms of usability for several theoretical and practical reasons. To conserve the semantic interoperability achieved with a centralized ontology, we present a technique for automated mapping between distributed ontologies. The mapping technique is based on a method for automated web page categorization.