Anomaly Detection In Cellular Network Data Using Big Data Analytics

Anomaly detection is a key component in which perturbations from a normal behavior suggests a misconfigured/ mismatched data in related systems. In this paper, we present a call detail record based anomaly detection method (CADM) that analyzes the users's calling activities and detects the abnormal behavior of user movements in a real cellular network. CADM is capable of detecting the location of the site that an anomaly has occurred. We evaluate the proposed CADM by performing experiments over the call-detail records of AVEA, a mobile service provider in Turkey.