Performance evaluation of anomaly detection in cellular core networks using self-organizing map

One of the preconditions to guarantee the quality of service (QoS) of cellular mobile networks is the rapid and accurate detection of key performance index (KPI) anomalies. This paper applies a neural network algorithm called self-organizing map (SOM) to monitor traffic measurement anomalies collected from an actual cellular network service provider. Results show that the SOM algorithm is able to detect global anomalies as well as identify which KPIs of the core network are abnormal. These results suggest that SOM can indeed help facilitate human operation, making it easier, faster and more efficient for human to troubleshoot, optimize or correct the configuration of the core network.