Personalized Multimedia Recommendation with Social Tags and Context Awareness

 Abstract—The large amount of multimedia contents often cause the problem of information overload. To tackle this problem, it is necessary to develop personalization techniques to recommend most suitable contents to users. In this work, we develop a new social tag-based method for the recommendation of multimedia items, and compare it with several often-used methods. A context-aware platform is also implemented that takes into account different environment situations in order to make the most sensible recommendations. It has generally been agreed that the collaboration-based approach can provide better performance than the content-based approach. Yet, with the current trend of organizing and sharing digital content through user-created metadata (i.e., social tags), the performance and effectiveness of collaborative recommendation can be improved by using such metadata to further recognize how the users likes specific items. Social tags are brief descriptions of items and they are freely supplied by a community of internet users to aid the access of large collections of media (5)(6). The use of social tags provides an interesting alternative to current efforts at semantic web ontologies in content annotation (7) (8). As tagging is neither exclusive nor hierarchical and therefore can in some circumstances have an advantage over hierarchical taxonomies. In this work, we adopt this way for multimedia annotation, use the tag information to analyze how the user likes specific items, and exploit such user