Improved Recommendation System with Friends on SNS

With the rapid development of the Internet, SNS services and 3G commercial mobile applications which brings tremendous opportunity, although the time on the development of SNS is very short in China, social web game is in the early stage of development, because of massive user, the potential commercial value of Chinese SNS is still a great mining space. A relatively large defects is the precipitation and accumulation on content, the dynamic of friends will affect our own decisions largely, it is favorable for the activity of SNS to increase the number of friends. We study some algorithms and models of the recommended system in this paper, and add the context awareness into the SNS friend recommended system, improve the calculation formula of similarity. We have improved the existing models, and conduct experiments to validate it and compare it with previous methods.