Research Of IPTV Hybrid Recommendation Method Based On Cloud Computing

With the continuous development of the Internet, more and more people came into contact with an Interactive Protocol TV(IPTV). IPTV user can use his own favorite video-on-demand programs and channels. However, when the user enjoy high-quality video service, they also faced with information overload and information Trek problems. Faced with the large number of users of the program, it is difficult to quickly find favorite shows. Therefore, how to help users quickly find favorite programs according to the user’s interest, has important significance.Personalized recommendation system is the most effective tool to solve the information overload problem today, good policies can be recommended for its recommendation based on user interest may meet user interest programs. However, IPTV system, user behavior and user behavior traditional recommendation systems under consideration are different, because TV is a user-oriented home entertainment devices, IPTV system is not a separate entity, but a family member has a different family different types of interest. For the traditional collaborative filtering recommendation speaking, IPTV can not get direct user ratings, and according to user preferences User rating description is inaccurate. For content-based recommendation, the recommendation of the results of too rigid, the accuracy is not high. Therefore, the traditional recommendation strategy recommended in the IPTV system is not very suitable. And, in the face of massive data, traditional stand-alone recommendation algorithms can not meet the requirements of IPTV systems in terms of efficiency.Based on these issues, this paper proposes an IPTV hybrid recommendation algorithm based on cloud computing, and the main researches are as follows:Through the analysis of user behavior records, the introduction of time context information to recommend the system to calculate the implicit rating, combined with the results obtained recommend items based on collaborative filtering algorithm.Establish IPTV tag library program,describing the program characteristics, and using content-based recommendation strategy, get the recommended results.The two recommended strategies are mixed to obtain the recommended strategy based on hybrid implicit context information and content. The method uses a distributed algorithm to improve in order to enhance the efficiency of the algorithm.The experimental results indicate that cloud computing platform based on network TV hybrid recommendation algorithm can effectively improve the accuracy of IPTV program recommended in this paper, while greatly enhance the efficiency of recommendation.