A Novel Network Intrusion Detection Model using Residual Recurrent Neural Network with Improved Garter Snake-based Optimization Strategy

The data security is considered to be one of the severe difficulties in the emerging internet globe. The data created via the internet is exposed to multiple kinds of exploits and threats. The security approaches like "Intrusion Detection Systems (IDS)" are developed to identify the multiple kinds of attacks and vulnerabilities. Numerous "machine and deep learning" techniques are employed for generating the IDS. The rapid enhancement of computer connectivity and the important amount of computer-based developments enhance in the modern days so the difficulty of satisfying the cyber-city is rising continuously. It also demands an eminent protection mechanism for various cyber threats. Hence, identifying the attacks and the inconsistency in the computer network and enhancing the IDS that work is a powerful part of the cyber security. In this task, a new network "intrusion detection" model is developed to effectively identify the malicious activity of unauthorized users to protect the individual’s computer systems. It helps to reduce the network traffic. The data is accumulated from the internet. The data is subjected to the optimal feature selection phase. Here, the suggested "Enhanced Garter Snake Optimization (EGSO)" algorithm is employed to select the best features. Further, the selected features are forwarded to the detection stage. The intrusion detection is carried out utilizing a Residual Recurrent Neural Network (RRNN) network. The implemented network "intrusion detection" model functionality is contrasted with numerous heuristic algorithms and methods.