Mathematical Models of Heterogeneous Machine Learning Techniques for Ransomware Protection in Cyber-Physical Systems

Ransomware attacks are a growing threat to cyber-physical systems (CPS), capable of causing financial losses, downtime, and even physical harm. Using machine learning to detect and prevent these attacks is a promising approach. A study compared four machine learning techniques—support vector machines (SVM), decision trees, random forests, and neural networks—to detect ransomware attacks on CPS. The models were trained and evaluated using a dataset of real ransomware incidents. Results showed that all four techniques effectively detected ransomware attacks. Neural networks achieved the highest F1 score, while SVMs excelled in precision, recall, and accuracy. However, the differences in performance among the methods were not significant. The models detected ransomware outbreaks within an average of 12 hours, allowing businesses to mitigate the attack's effects promptly. The study concluded that machine learning could effectively detect and prevent ransomware attacks on CPS. However, it is important to note that the study used a limited dataset, highlighting the need for further research with larger datasets to confirm these findings.