A concept for co-existence of heterogeneous recommender systems based on blackboard architecture

Abstract

This paper introduces the concept of software coordination, collaboration, moderation and visualization in the context of Recommender System (RS). The proposed concept assist to coordinates to execute multiple RSs and collaborates their respective outputs. It also moderates the top-n recommendation list generated by RSs. The multiple RSs can be heterogeneous in terms of their attributes, such as data mining algorithms & prediction techniques used for recommendation. These RSs may be developed by independent experts. The concept employs the notion of blackboard architecture for dealing with multiple RSs. Optimal results are expected to be achieved by feed-forward and feed-backward mechanism.