Classification and Ontology Maintenance in Agent-Based Knowledge Management Frameworks: A Prototypical Approach

Being able to create views on the document space via grouping the documents is a key functionality in intelligent document management in view of browsing and querying. Hierarchically grouped sets of Documents can be viewed as simple extensionally defined ontological concepts. In an example Knowledge Management system (KnowCat) developed at UAM, Madrid, we investigate how agents for the maintenance of this ontology (these document groupings) can be constructed. We discuss two examples: A classification agent and a maintenance agent support users and administrators of the system to keep the ontology tight and functional. The agents are tested, developed targeted toward Spanish natural language documents, which requires adapted NLP techniques.