A study about the personalized recommendation system of TV program based on user behaviors

Human behavior can be a direct reflection of their interest and purposes; therefore, we can take advantage of user behavior to predict their interest in the recommendation system. This paper analyzed the relationship between user behavior and interest in the recommender system of personalized video program, and summarized three potentially useful observations: selection, duration, repetition. Implicit feedback, inferred on the three behaviors, is introduced into the system. Then the paper presented a method of measuring user interest from implicit feedback, based on which constructed user preferences model by combining all implicit feedback.