Benchmarking Unsupervised Machine Learning for Mobile Network Anomaly Detection

Abstract

A mobile network is a type of communication system that reaches end-users with high speed, availability and provides many services without any delay. The type of services is in the health sector, emerging market, emergency services, and many more. To ensure service availability mobile network monitoring in now gone to another level by introducing many automated ways, introducing advanced analytic. In a mobile network, multiple nodes are interconnected to each other. Those nodes are spread across various network domains and often produce a data type like network performance data. Over this huge data set, machine learning is useful to detect an anomaly for a different domain. But implementing supervised machine learning over large data will be difficult to use because it requires ground truth determination. On the other hand, unsupervised machine learning will cluster various data volumes without any prerequisite. If proper tuning is imposed on this model, it will give an outcome of efficient anomaly detection. The key content of this paper is to identify unsupervised machine learning which is best suited for mobile network anomaly detection. To do that a benchmark approach is performed over three unsupervised machine learning and these are K-means, DBSCAN, and HDBSCAN. The thumb rule of the benchmark is followed as converting the unsupervised machine learning output into a classification problem and then measuring the model performance.