摘要

This paper propose a framework Knowledge advantage machine (KAM) to help in organizing individually discovered knowledge drawn from a narrowly bounded domain into a personal knowledge network based on personal request and tags. Ontologies, folksonomy and personomy are employed in KAM to constitute the useful repositories of knowledge. Ontologies offer a flexible and expressive layer of abstraction, very useful for capturing the semantics of information repositories, but they can not reflect the user's interest. The user in KAM can freely choose the words to tag the resources which are the reflection of the user's own interest. The set of tags and tagged knowledge of a user comprise the personomy. In a group the shared tags and knowledge are known as folksonomy. Our approach investigates how to map these tags in personomy and folksonomy to existing domain ontology in order to add accurate meanings. The user's behaviors are also used to re-rank the query results. So the user can find the useful knowledge quickly and accurately.