TV program recommendation using implicit feedback with adaptive regularization

Most collaborative filtering methods use explicit ratings given by users for some items. On the other hand, as in TV program recommendation, sometimes it is not easy to obtain those explicit ratings. Instead of ratings, there is information on how long the user watched an item. For such problems, instead of explicit recommendation methods, implicit methods should be used. In this paper we process the time durations for which users watch the programs to obtain rankings and also use the normalized time durations for TV program recommendation. As the recommendation methods, we use explicit recommendation methods together with matrix factorization and relate these methods with the implicit recommendation methods. While learning the matrix factors, we introduce adaptive learning rate to speed up the learning and we also introduce user/item adaptive regularization.