METHOD, DEVICE, AND PROGRAM FOR CALCULATING VIEWING PROBABILITY, AND METHOD, DEVICE, AND PROGRAM FOR RECOMMENDING VIEWING PLAN

摘要

PROBLEM TO BE SOLVED: To recommend a video matching recently viewed videos and a user's interest in accordance with viewing history information and to enable the user to be provided with a viewing plan.

SOLUTION: Markov model generation means acquires viewers' viewing histories from viewing history storage means and calculates viewing probabilities of a certain video on the basis of a Markov model. Topic model generation means acquires viewers' viewing histories from the viewing history storage means and calculates viewing probabilities of the certain video on the basis of a topic model. Viewer action model generation means repeatedly executes processing of obtaining a probability that a certain viewer having a certain viewing history views the certain video, by binary operation of probabilities calculated by the Markov model generation means and the topic model generation means and obtains a probability that the certain viewer having the certain viewing history selects a certain viewing plan.

COPYRIGHT: (C)2012,JPO&INPIT