抄録

I-036
Personalized TV-program Recommendation using User's Viewing Histories and Time-Frames

We propose a new recommendation algorithm called “time-frame watch-flow algorithm” for selecting K highest potential TV programs that a user might like based on the user’s viewing history and a specific time-frame during a day. Based on our method, the recommending value is assigned for each TV program based on the hypothesis that users’ preferences are dynamically changing by time-frame in a day. Furthermore, the proposed method is also capable of giving a personalized recommendation for a specific user based on his/her watching sequence, improving the prediction accuracy and the diversity.