5G RAN service classification using Long Short Term Memory Neural Network

5G brings many benefits such as enlarged capacity and improved connectivity. However, it also poses challenges especially due to a significant increase in the amount of traffic on the network. This creates difficulties for operators to maintain the Quality of Service (QoS) for each of the services offered. Therefore, in order to improve the performance of such capabilities and, consequently, the experience of the users, it is necessary to identify which traffic requires more prioritisation. This would help allocate more resources to those services. This concept makes the identification and classification of traffic to gain more and more relevance and importance. In this paper, we propose a Long Short-Term Memory (LSTM) model to classify 5G Radio Access Network (RAN) behaviour into four different scenarios: streaming, video conferencing, Voice over IP (VoIP) and gaming. The results obtained show a $93 \%$ accuracy.