Prediction of user traffic in cellular networks has attracted profound\nattention for improving resource utilization. In this paper, we study the\nproblem of network traffic traffic prediction and classification by employing\nstandard machine learning and statistical learning time series prediction\nmethods, including long short-term memory (LSTM) and autoregressive integrated\nmoving average (ARIMA), respectively. We present an extensive experimental\nevaluation of the designed tools over a real network traffic dataset. Within\nthis analysis, we explore the impact of different parameters to the\neffectiveness of the predictions. We further extend our analysis to the problem\nof network traffic classification and prediction of traffic bursts. The\nresults, on the one hand, demonstrate superior performance of LSTM over ARIMA\nin general, especially when the length of the training time series is high\nenough, and it is augmented by a wisely-selected set of features. On the other\nhand, the results shed light on the circumstances in which, ARIMA performs\nclose to the optimal with lower complexity.\n