Knowledge based personalization

Abstract

Semantic Web aims at bringing order to the information chaos on the web by associating semantics and context with content, thereby improving its value. It focuses on making content usable by computers for answering questions and performing tasks on behalf of users. Knowledge sharing using ontologies is critical for realizing this vision. Users would create and manage ontologies and share them on a network, thus creating a personalized view of the knowledge and still interoperating with other ontologies in the network. Searching for ontologies that best fit a user.s perception is non-trivial. To help him do this, his personalized knowledge base, profile and query history can be used to deploy personalization techniques. We discuss this knowledge based personalization approach as used in the InfoQuilt system. We describe the use of multiple personalization techniques used in combination and particularly, query relationship to predict related ontologies based on current context. Query relationships use InfoQuilt.s framework for inter-ontological relationships to do this by describing relationships between queries on two domains.