Intrusion Detection for Cyber Physical Systems Using Light Gradient Boost Model

Today’s massive internet connectivity drives increase in cyber physical infrastructure necessitating attention on random network attacks mitigating techniques. Military and defenses cyber systems particularly stand to lose more by the cyber-attacks, where the absence of mitigation could result in comprise of weapons blueprints, operational schemes etc., and seriously threatening national security. So, mitigating frameworks utilized to shield the systems from network dangers and guaranteeing elevated degree of safety has gained significant attention. Moreover, end users are still adapting to using the cloud storage where information must be securely transferred by shielding the infrastructure from intrusion attacks. The proposed system is focused on creating a robust model to detect network attacks coming as intrusion for Internet of Thing (IoT) devices. The system develops a LGB regression model using CICIDS2018 dataset. The presented approach considers various attributes as key whole parameter for finding the presence of intrusion attacks over the network. The presented system achieved MSE 0.8670 and compared with various states of art approaches and processing delay of 17.24 s is achieved.