Bayesian model for assessing cybersecurity risks based on Naive Bayes classifier

The research analyzed cybersecurity data and interpreted the relationship between two variables. The first variable, the dependent variable, is called the "risk level," which is classified into two categories: high and medium. The second variable, the independent variable, comprised three sub-variables. The first sub-variable, called the "type of cyberattack," included eight main types identified based on the most common cybersecurity threats: (phishing email attacks, malware insertion via attachment, Distributed Denial of Service (DDoS) attacks, SQL injection attacks, denial-of-service attacks on main servers, email or phone calls to steal accounts, and viruses spread via malicious links). The second sub-variable, called the "impact of the attack," included four types: (data loss, data disruption, service disruption, and data leakage). The third sub-variable, called the "security level," was divided into (low, medium, and high). These sub-variables were obtained by distributing a questionnaire to 175 cybersecurity professionals, which included questions about the perceived impact of breaches. The study examined cybersecurity database attacks and how to protect against them, analyzing them using a Bayesian classifier and the R statistical programming language. The results showed clear differences in the level of risk depending on the type of attack.