Recommendation system of IPTV TV program using ontology and K-means clustering

초록

In this paper we introduce a recommendation system for recommending preferred TV genres for each viewer, using ontology technique and K-means clustering algorithm. The algorithm is developed based on the personal VOD viewing history. First, the viewing history is built in an ontology which is able to achieve inference process through a query. In the list of users, each item and class obtain the probability of preference of VODs and then the information is used for building the ontology. From the ontology we select each user's preferred VODs using K-means algorithm. In the experimental section, we show the feasibility of our algorithm using real TV viewing data. © 2011 Springer-Verlag.