A recommendation system based on object of the interest

Needs of contents recommendation technologies are increasing as the number of the media content on TV and Internet is increased. For the precise recommendation, researches on collaborative filtering recommendation system are increasing. But these researches have problems in scalability and cold-start. To solve these problems, this paper proposes the recommendation system based on user's content consumption according to watching time pattern and objects of interest. For this, we extract objects of interest from 259 media contents and 157 users and then calculate preference scores. To evaluate the performance of the proposed system comparison tests are done with the results of the existing collaborative filtering systems.