Knowledge flow leveraged through case-based and data mining agents in a just in time information and knowledge system

摘要

Information and knowledge flow is the central component in the architecture of a knowledge management (KM) system. Several architectures of KM systems can be found in the computer literature. In many architectures the knowledge flow is not clearly established. JITIK (for Just in Time Information and Knowledge) is a distributed agent-based system whose goal is to transport the right knowledge to the right people at the right time. In this paper we propose the integration to JITIK of service agents that use knowledge technologies in order to detect and distribute knowledge to people in an organization. We are interested in two mature technologies: case-based systems and data mining. Agent services for these technologies were specified, implemented and integrated to JITIK. The agents were tested with several use cases. The integrations proposed represent a method to automatically exploit knowledge sources using specific knowledge technologies in KM flow systems. It also represents the first agent services in JITIK that clearly exploit knowledge sources.