Cellular traffic type recognition and prediction

4G and 5G cellular traffic pattern recognition and prediction are key objectives for network optimization. They also are becoming of fundamental importance for the next-generation cellular network. Recognizing mobile traffic patterns and proactively knowing the user behaviors allow the operator to optimize the resource allocation. On the other hand, it is a complex problem due to the diverse set of applications behind the traffic. Most traffic prediction problems focus on capturing the dynamic of traffic and enhancing the performance. In this paper, we design a deep learning model for traffic pattern recognition and prediction of the type of arrival packet using Long Short-Term Memory (LSTM) neural networks. The mobile traffic information is collected from the Downlink Control Information (DCI) using the Amarisoft software. The learning phase of the model relies on a well-known traffic pattern simulated on Amarisoft 4G and 5G testbed.