A Constrained Spreading Activation Approach to Collaborative Filtering

In this paper, we describe a collaborative filtering approach that uses features of users and items to better represent the problem space and to provide better recommendations to users. Features of the collaborative filtering dataset are found and incorporated into a network representation of the collaborative filtering space where users and items are represented by nodes and where the nodes are connected by weighted edges. A spreading activation approach to collaborative filtering, using this representation and constrained by the user and item feature information, is compared with a traditional collaborative filtering approach.