Mathematical Modeling of Cyberattack Defense Mechanism Using Hybrid Transfer Learning With Snow Ablation Optimization Algorithm in Critical Infrastructures

Cybersecurity is a significant topic that has turned into an efficient one at present owing to the increasing dependency on interconnected methods and technology. As digitalization upsurges, the requirement for cybersecurity measures becomes even more vital for numerous networks. To certify the security of critical infrastructures, numerous cyber security solutions must be taken together and the essential infrastructure must be developed. Industrial control methods are one of the most vital aspects of the cybersecurity of critical infrastructures. It is possible to entirely stop these physically located devices’ operations and do substantial destruction with a cyberattack. Whereas AI solutions are being utilized in numerous fields, cyber security has started to become one of the concentrated fields of artificial intelligence (AI) domain. Consequently, there are many studies on identifying cyberattacks by utilizing AI methods. It is probable to utilize AI to help and support cybersecurity solutions to develop cybersecurity of significant infrastructures. This study develops a Cyberattack Defense Mechanism using Hybrid Transfer Learning with Snow Ablation Optimization Algorithm (CDMHTL-SAOA) technique in Critical infrastructures. The main cause of the CDMHTL-SAOA model is to improve the cyber security maturity level of critical infrastructures and inspect both traditional cyberattack and AI approaches. Primarily, the data normalization process can be implemented to scale the raw data into a uniform format. In addition, the snow ablation optimization (SAO) algorithm can be exploited for the optimum choice of feature subsets. For the cybersecurity classification process, the presented CDMHTL-SAOA technique applies the hybrid of convolutional neural network and bi-directional long short-term memory (CNN-BiLSTM) method. Eventually, the parameter choice of the CNN-BiLSTM technique has been implemented by the design of the hippopotamus optimization algorithm (HOA). To represent the better solution of the CDMHTL-SAOA classifier, a simulation validation can be tested on a benchmark database and the solutions are measured for various aspects. The simulation outcomes certified the improved execution of the CDMHTL-SAOA method over other techniques.