Location Based Content Recommendation for CASoRT System

The current wireless communication network is a content-careless network, in which every request needs a new transmission, even those requests imply the same content. However, the Long-Tail distribution of users' interests caused by the converging behavior, determines that the redundant traffic, which is the re-transmission bearing same content, is rather enormous. Therefore the content-careless scheme of network transmission introduces large energy waste. In our previous work, we put forward a novel transmission scheme CASoRT to mitigate such energy waste by broadcasting those hot contents. After the contents are broadcast, the terminals in coverage could cache those contents in local memory and load those contents when request. Thus, the massive redundant transmission of the very content is relieved by one broadcast transmission. In order to exploit the advantage of broadcast, the proper recommender scheme, deciding which contents should be broadcast, needs to be optimized. In this paper, we present certain perspective in analyzing the log of campus network in Tsinghua University. It shows that utilizing users converging character could improve the efficiency of CASoRT with analysis. Finally, location based content recommendation scheme is proposed, verified by simulation.