Research Of Mobile Traffic Prediction Based On Gray Model

With the application of mobile communication networks broaden and increased in size, the importance of network management is highlighted. A new generation of network management systems in the field of traffic forecasts also have a new demand in forecasting traffic analysis has become an extremely important part. Accurate traffic forecast for network management, planning, design is of great significance. The base station capacity in recent years increasingly saturated, slowing growth in the number of base stations, operators have to face the problem of not only the saturation of network resources in some areas, even in the face to reduce the situation.The problems not only include limited capacity of base station, but also the difficulties of expanding the number of base station, and even facing the decreasing of base station. How to provide more services to users with limited resources on the physical carrier frequency and base stations, while ensuring quality of service to further enhance and become an important issue facing the mobile communication network. Therefore subject to traffic forecasts for communications operators more and more attention. In the daily operation and management of the communications industry, the traffic forecast is still not a mature fields. The accurate prediction is directly related to the corporate network quality and user perception.Mainly, forecasting is based on historical data and experience guiding drawing traffic trends. The prediction methods decide the result of prediction; the choice of method of forecasting is the core and difficult job. This paper aims to mobile communication network traffic with time-varying nature of an analysis, the establishment of an appropriate prediction model research. Including traffic cycle component of time series analysis, trend analysis and other component is the focus of research. GM(1,1) model,as one of the most important part of the gray system theory is becoming more valueable in expanding fields. The outstanding features of the GM model including simple modeling,briefly representing, flexible applying. In dealing with the core problem of trend analysis, GM model is used this paper and the procedur is displayed. By using the methords like residual test, the model is proved capable to be used in predicting traffic. Telecommunications operators may grasp the real-time user behavior, traffic trend of development and carry out rational decision-making and investment.