User Traffic Prediction Based on K Neighbors Collaborative Filtering for CASoRT System

Repeat transmission of hotspot traffics results in great waste of energy and bandwidth in wireless network, for the communication of current network is content careless. To address this issue, a novel content aware transmission schema named CASoRT System was put forward in our previous research to reduce wireless resource waste by the benefit of broadcast. In this paper, we propose a K neighbors collaborative filtering prediction method to forecast the request of hotspot traffics, which will be broadcasted by CASoRT. Firstly, we introduce the traditional prediction methods and describe the framework of our job. Then, we deliver the expression of our method in formulation, and discuss the time complexity in quantitative analysis. At last, our method is validated by simulation, and the results demonstrate that the new schema can potentially lead to 20% deduction compared with the unicast schema in hotspot traffic transmission. And the results also show that our method reduces 25% prediction time than traditional methods, but with the same performance in saving wireless resources.