Context-aware ontological schemes for multimedia personalization

Current methods used to explore the possibility of using context-aware ontological semantics for evaluating user-profile similarity are limited. The contributions of this paper are three-fold: a) to utilize and leverage the concept of ontology in p2p-based recommender system, b) to evaluate ontologically expanded user-profile for similarity evaluation, and c) to address the issue of cold-start in p2p-based personalization systems. We evaluated our method using user-profiles, extracted from Netflix Dataset, in measuring similarity between IPTV peer-users. Our method enriches user-profile vector for enhanced similarity evaluation with other peer-users in p2p network. Our approach expands the vector components for similarity evaluation with the other IPTV peer-users' expanded user-profile. This overall facilitates better comparison, better recommender aggregation, and quality recommendations.