Traffic Prediction Model for Telecommunication Base Stations Based on Improved WOA Optimized LSTM and EMD

A hybrid telecommunication base station traffic prediction model based on a multi-strategy whale optimization algorithm (MWOA) optimized long short-term memory (LSTM) neural network and empirical mode decomposition (EMD) techniques is proposed to achieve better communication base station traffic prediction performance. First, to ensure that the prediction model has sufficient extraction of the features of the base station traffic data to further improve the prediction accuracy, the non-smooth telecom base station traffic data is decomposed into intrinsic mode function (IMF) components and residual components using EMD. Second, the hyperparameters of the LSTM neural network are optimized by the MWOA optimization algorithm, which can avoid the blindness of artificial selection of hyperparameters and make the model converge to the best-fit state, thus obtaining more satisfactory prediction results. Finally, the prediction values of each component in the model are collected and processed as the prediction results of communication base station traffic. By comparing with several models, the EMD-MWOA-LSTM model has better performance.