Traffic Prediction Model of Power Wireless Private Network Based on LSTM Neural Network

With the development of ubiquitous Internet of Things technology, more and more power equipment terminals will be connected to the power wireless private network in the future. The network communication quality and network stability will face great challenges. The tension of spectrum resources will be further enhanced. This paper obtains the base station historical traffic data by monitoring the base station service traffic, and then predicts the future service traffic changes in the time, and then adjusts the resource allocation. This paper first analyzes the temporal and spatial characteristics of base station traffic to screen and eliminate the dynamic changes of individual base station traffic in a certain area, then clusters the base station traffic, cluster the base station in the network, and test the same cluster base station data together, build the corresponding cluster data set, combined with LSTM model training, using the maximum value pooling to evaluate the LSTM network model.