Improving Network-Based Top-N Recommendation with Background Knowledge from Linked Open Data

Abstract

The boom in Linked Open Data (LOD) has recently stimulated the research of a new generation of recommender systems—LOD-enabled recommender systems. ROUND (Random walk with restart on an Object-User Network towards personalized recommenDations) is a state-of-the-art method for network-based top-N recommendation. However, the ROUND method relies solely on the historical data (i.e., the ratings matrix) and does not take full advantage of background knowledge from LOD. This paper addresses the problem of improving network-based top-N recommendation using background knowledge from LOD by proposing an improved ROUND method called ROUND-APICSS. The core idea of ROUND-APICSS is that we exploit a knowledge graph constructed from LOD to calculate semantic similarities between the objects (items) involved in the recommender system, thereby improving the object-user heterogeneous network model and the random walk with restart model on the network. Our experimental results on real datasets suggest that the incorporation of background knowledge from LOD into the network-based top-N recommendation models can improve recommendation accuracy. The results also show the superiority of our ROUND-APICSS method over the ROUND method in terms of recommendation accuracy.