Personalized Movie Recommendation Based on Social Tagging Systems

In the Internet era, the movie website has become a mainstream platform for movies introduction and comment on the movie resources, annotation also has become a mainstream form of movie resources under the network environment of the new organization based on social tagging.. From the system point of view, the greatest advantage offered by tagging applications is the richness of the tag profiles. However, the freedom afforded users comes at a cost: an uncontrolled vocabulary can result in tag ambiguity hindering navigation. Thus, a key question is how to harvest tag semantics from these systems. We present an algorithm of tags clustering. With this algorithm, we clustering the tags into a semantic tree, then we turn every movie item into an induced tree. We propose a new method for movie recommendation, which based on semantic similarity of tags. When we retrieve the movie based on semantic similarity of tags, our algorithm shows the high precision.