A recommendation algorithm for Knowledge Objects based on a trust model

This paper presents a recommendation algorithm with which to recommend Knowledge Objects within a Community of Practice (CoP). It is based on a trust model that takes into account not only previous experience of a Knowledge Source, but also the position of the respective member in the community, along with that member's level of expertise. Furthermore, the system attempts to emulate the principle of human intuition. The combination of these factors, which can be adjusted by means of weight factors, enables the system to make decisions based on a calculated trust value even if a new member is introduced into the community or the entire community is newly created. To demonstrate the capabilities of this algorithm, a tool employing a multi-agent architecture has been developed which is also presented.