Comparative Study of Learning Effectiveness Across NEPs on Network Fault Correlation

The paper presents a comparative analysis of performance of various network equipment providers (NEPs) operating radio access network (RAN) by fault correlation using error code. Effectiveness is measured using supervisory learning algorithms applied on network alarm data derived from fault management system (FMS) along with key performance indicators (KPIs) derived from performance management system (PMS). Results are based on the ML reinforcement model algorithm is tested with live data stream for one large CSP in India with accuracy of ~ 90% using vendor [Ericsson/Nokia/Huawei] and technology [2G/3G].