Short term prediction of wireless traffic based on tensor decomposition and recurrent neural network

Abstract This paper proposes a wireless network traffic prediction model based on Bayesian Gaussian tensor decomposition and recurrent neural with rectified linear unit (BGCP-RNN-ReLU model), which can effectively predict the changes in upstream downstream short period of time future. The research is divided into two parts: (i) missing observations are imputed by an algorithm decomposition. (ii) used to forecast true only rather than both estimated observations. results show that, compared other combined models data imputation networks, BGCP-RNN-ReLU proposed this has smallest error for traffic. new achieves better forecasting precision, thus help regulate load communication station reduce resource consumption. Highlights problem values stages handle. A newly propose d method more efficiently impute data. Simple obtains performance complex networks.