A mixture model for multi-source cyber-vulnerability assessment

Cyber-risk assessment is gaining increasing attention due to the potential impact of cyber-incidents on digital and physical systems, with consequent effects on individuals and organisations. This work introduces a Bayesian mixture model to relate individual cybervulnerability features and accessible information that could affect attackers’ evaluation and cyber-incident occurrence. Individual cybervulnerability features are clustered according to a Dirichlet process. Future research directions aimed at prioritising cyber-vulnerabilities and access to information are discussed too.