Network traffic analysis and prediction of Hotspot in cellular network

Nowadays, mobile devices become more and more important in people's daily life. A lot of mobile devices connect to the internet through the base stations. Since base station has limited resources such as signaling and spectrum, it is a challenge for network operator to manage network properly. A base station with high traffic amount and large number of subscribers, which can be considered as hotspot, may provide low quality of service(QoS). In this paper, we focus on four important features which will influence the QoS and propose an entropy weight based method to identify the hotspot. Then we analyze the spatial-temporal distribution of hotspot and identify the hot region. At last, we use ARMA and Holt-Winters to predict the traffic amount of hotspot. Our results will help network operator to manage network resource properly and improve QoS in hot region.