Browse by author
Lookup NU author(s): Dr Ayon MukherjeeORCiD, Dr Manuel HerreraORCiD, Aman Parekh
This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0).
Telecommunication base station monitoring systems trigger alarms in response to network events, degradations, and recoveries. These alarm sequences are crucial for predictive analytics and quantification of resilience not only in the base stations, but the entire network. These alarms reflect structured information flows that can reveal the dynamic state and interdependencies of network components. This work proposes RAPTOR, a framework that integrates temporal clustering, first-order Markov modeling, and resilience metrics to analyse and probabilistically forecast alarm behaviors across Radio Access Networks (RANs). Using Dynamic Time Warping (DTW) to identify co-activated alarm clusters and Markov transition matrices to estimate alarm propagation, the proposed RAPTOR framework supports accurate modelling of information flow, next-alarm prediction, and interpretable resilience scores based on alarm entropy, self-loop tendency, and absorption probability. The approach is developed and validated on real-world alarm logs from multiple base stations from a major telecommunication service provider in the UK. The proposed framework demonstrates strong prediction accuracy and the ability to reveal operational complexity and network fragility in a structured and adaptive manner.
Author(s): Mukherjee A, Chandra A, Herrera M, Li L, Indiran H, Parekh A, Parlikad AK
Publication type: Article
Publication status: Published
Journal: Journal of Risk and Reliability
Year: 2026
Pages: epub ahead of print
Online publication date: 26/03/2026
Acceptance date: 09/02/2026
Date deposited: 25/02/2026
ISSN (print): 1748-006X
ISSN (electronic): 1748-0078
Publisher: Sage
URL: https://doi.org/10.1177/1748006X261431895
DOI: 10.1177/1748006X261431895
Data Access Statement: The data underlying this study are proprietary to the collaborating industry partner and cannot be shared publicly due to confidentiality and contractual restrictions. Derived, de-identified, or aggregated results may be available from the corresponding author upon reasonable request and subject to the partner’s approval
Altmetrics provided by Altmetric
