Cell base station traffic prediction based on GRU

Abstract

With the expansion of Internet technology and network scale, the data volume of base station traffic also shows explosive growth. Predicting base station network traffic has high practical guiding significance for network research, management and control. Aiming at the problem of accurate prediction of base station traffic, this paper proposes a gated recurrent unit neural network model (GRU model) based on neural network algorithm, which can predict the base station traffic data according to the periodicity and fluctuating characteristics of base station traffic data. After experimental verification, it shows that compared with the traditional time series prediction model AR model, ARIMA model also has the convolutional neural network model based on neural network algorithm. This method has higher accuracy and smaller experimental error in mobile communication traffic prediction. The MAE value is optimized by 27.04%, 37.89% and 9.12%.