Developing Cybersecurity Systems Based on Machine Learning and Deep Learning Algorithms for Protecting Food Security Systems: Industrial Control Systems

Industrial control systems (ICSs) for critical infrastructure are extensively utilized to provide the fundamental functions of society and frequently employed in infrastructure. Therefore, security these from cyberattacks is essential. Over years, several proposals have been made various types cyberattack detection systems, with each concept using a distinct set processes methodologies. However, there substantial void literature regarding approaches detecting ICSs. Identifying ICSs primary aim this proposed research. Anomaly based on an artificial intelligence algorithm presented. The methodology intended serve as guideline future research area. On one hand, machine learning includes logistic regression, k-nearest neighbors (KNN), linear discriminant analysis (LDA), decision tree (DT) algorithms, deep long short-term memory (LSTM), convolution neural network (CNN-LSTM) detect ICS malicious attacks. algorithms were examined real datasets industrial partners Necon Automation International Islamic University Malaysia (IIUM). There three attacks: man-in-the-middle (mitm) attack, web-server access telnet well normal. system was developed two stages: binary classification multiclass classification. detected malware normal or attacks used all individual KNN DT achieved superior accuracy (100%) Moreover, sensitivity method presented predict error between target prediction values. results showed that R2 = 100% both stages. obtained compared existing systems; outperformed systems.