Long short-term memory neural network for network traffic prediction

This paper proposes a model of neural network which can be used to combine Long Short Term Memory networks (LSTM) with Deep Neural Networks (DNN). Autocorrelation coefficient is added to model to improve the accuracy of prediction model. It can provide better than the other traditional precision of the model. And after considering the autocorrelation features, the neural network of LSTM and DNN has certain advantages in the accuracy of the large granularity data sets. Several experiments were held using real-world data to show effectivity of LSTM model and accuracy were improve with autocorrelation considered.