Correlation and Dependence Analysis on Cyberthreat Alerts

Printer-friendly versionSend by email
John M.A. Bothos, Konstantinos-Georgios Thanos, Dimitris M. Kyriazanos, George Vardoulias, Andreas Zalonis, Eirini Papadopoulou, Yannis Corovesis, Stelios C.A. Thomopoulos, National Centre for Scientific Research “Demokritos” (NCSRD), Greece
In this paper, a methodology for the enhancement of computer networks’ cyber-defense is presented. Using a time-series dataset, drawn for a 60-day period and for 12 hours per day and depicting the occurrences of cyberthreat alerts at hourly intervals, the correlation and dependency coefficients that occur in an organization’s network between different types of cyberthreat alerts are determined. Certain mathematical methods like the Spearman correlation coefficient and the Poisson regression stochastic model are used. For certain types of cyberthreat alerts, results show a significant positive correlation and dependence between them. The analysis methodology presented could help the administrative and IT managers of an organization to implement organizational policies for cybersecurity.
Integrated Systems Laboratory (ISL)
Publication Name: 
ITU Journal: ICT Discoveries

© 2019 - Institute of Informatics and Telecommunications | National Centre for Scientific Research "Demokritos"

Terms of Service and Privacy Policy