AN ADAPTIVE AGENT-BASED FRAMEWORK FOR KNOWLEDGE MANAGEMENT AND SHARING
Jinsoo Park, Dongwon Lee · 2001
Abstract This paper presents an adaptive agent-based framework for knowledge sharing. The framework supports semantic-based knowledge retrieval and filtering through the usage of ontologies, case-based reasoning, and genetic algorithms. First, the concept of ontologies is applied to build users’ profiles and resolve semantic conflicts among multiple knowledge sources – User Profile Ontology (UPROL), Semantic Conflict Resolution Ontology (SCROL), and Domain ontology. UPROL is used by the profile agent to map users’ interests into a common terminology. SCROL allows multiple views and interpretations of a given terminology by different users and applications. Domain ontology is also used to organize information sources and direct search processes. We also utilize collaborative filtering through case-based reasoning techniques for knowledge filtering and recommendation. Finally, by adopting the survival-of-the-fittest rule of genetic algorithms, we make the system continuously adapt to changing users’ profiles. In conclusion, our adaptive agent-based approach allows knowledge seekers to access tacit yet interoperable knowledge.