Forecast Of Mobile Network Traffic And The Number Of 5G Users

With the promotion and popularization of 5G technology,People's Daily work and life more need mobile network as the basis,and the performance of mobile network is increasingly demanding.Therefore,the prediction of mobile network traffic can provide a basis for abnormal network monitoring,operation and maintenance management,and then improve the overall network performance.At the same time,the number of 5G users continues to grow.Forecasting the data volume of 5G users in the next period of time can provide data basis for 5G network operators and terminal equipment manufacturers to formulate production and operation strategies.Aiming at mobile network traffic prediction,this paper first analyzes the characteristics of daily changes of mobile network traffic,and builds a comprehensive and effective prediction model based on the characteristics of network traffic.Through the analysis of the characteristics of network traffic change,it is found that the network traffic has such characteristics as self-similarity,periodicity and suddenness.The existence of these characteristics is also the basis to predict the future data value according to the historical data of some network traffic.With the adjustment of the two parameters of the wavelet function,the wavelet neural network can not only analyze the characteristics of the whole signal,but also focus on the changes of the local signal.It is especially suitable for dealing with sudden signals and non-stationary random signals.Therefore,the wavelet neural network is used for experimental prediction.In this paper,the hierarchical structure of the model is designed at first,and then the optimal number of hidden layer nodes is determined by multiple experiments.Finally,the wavelet model is established for simulation experiments.The experimental results show that this model can fit the changes of data well,but its prediction accuracy is not high,and the convergence speed is slow.Based on the results,an improved algorithm is proposed to optimize the ability of wavelet neural network to find the best parameters,and overcome the problem that it is easy to find the local optimal parameters.In this paper,the particle swarm optimization algorithm is used to improve the parameter adjustment process of the wavelet neural network.The particle swarm optimization algorithm can adjust the range and ability of searching the optimal solution,so it can solve the problems of the wavelet neural network prediction.Based on the flow prediction,the fitness function and particle vector composition of the particle swarm optimization algorithm are set,and then the hierarchical structure,optimal population number and other parameters of the model are designed to establish the model for prediction.Finally,the experimental results show that the prediction accuracy of this model is higher and the network training convergence is faster,which proves the effectiveness of the particle swarm optimization wavelet neural network model.The basic idea for predicting the number of 5G users is to regard the diffusion of 5G users as the change of adopters after a new product is put into the market.The most effective mathematical model in the field of predicting the diffusion of new products is Bass model.However,Bass model is more used to predict the diffusion of physical innovative products,and less to predict the products of 5G communication technology.Therefore,this paper analyzes the diffusion path and factors of 5G users,and then uses this model to fit the diffusion of 4G users,proving the applicability and effectiveness of Bass model in predicting the diffusion of 5G users.Finally,the historical data of the number of 5G users are counted,and the parameters are obtained by fitting,and the number of users in the next two years is predicted.