Tuning Personalized PageRank for Semantics-Aware Recommendations Based on Linked Open Data (original) (raw)

The paper discusses the impact of Linked Open Data (LOD) on graph-based recommendation systems, particularly through the tuning of the Personalized PageRank algorithm. By employing feature selection techniques, the methodology aims to optimize recommendation efficacy in the context of LOD. Experiments conducted demonstrate that integrating LOD features significantly improves recommendation performance across various datasets, including MovieLens and DBbook, indicating the effectiveness of the proposed approach.