A recommendation algorithm for IPTV video-on-demand system

摘要

As a new generation of digital cable TV products, the number of IPTV users is growing rapidly in China nowadays due to its good interaction. But the studies of recommendation systems are mainly for e-commerce which is different from IPTV, so it is even harder to find out which programs the users are interested in. To solve this problem, we analyze the difference between e-commerce and IPTV, and present a new recommendation algorithm for IPTV video-on-demand system. This paper uses the datasets of iTV programs-on-demand from Fujian telecom system to validate the effectiveness of the proposed experimental algorithm. The result of the experiment shows that when compared with other algorithms, the recommendation algorithm we proposed has the better F-measure, which means that the prediction accuracy and performance has been improved.