A semantic content-based recommender system using bayesian networks

In the era of digital world and WWW, most of the human activities have slowly started to be tightly coupled to the Internet. Like all other forms of multimedia web content, the amount of video content on the web has increased drastically over the past decade, reinforcing the need for Recommender Systems to help users reach relevant and interesting content. In an attempt to extend the research in the field of Recommender Systems by introducing cutting edge technologies, this thesis proposes a new recommendation approach in which Bayesian Networks are used for semantic aware reasoning about users interests. The theoretical proposal is accompanied by an illustrative implementation which is evaluated to verify the applicability of this approach. The results show that the proposed approach shows a very promising potential in practical applications with some limitations that could be further investigated for improvement.