Enhanced Intrusion Detection based on Advanced Deep Scaled Multi Perceptron Neural Network with Enhanced Particle Swarm Intelligence
N. Nathiya, S Mohanarengan, Diwakar S, K Chinnamannan, D Pugazhraj, Elango S T · 2024
Network communication is an incredible service for sharing the information among all over the world. Among the communication, the data breaches are occurring during transmission is increasing due to hackers plays to attack the service. By the intrusion created by the attackers during the commutation, through cyber-attack become more powerful to create various attacks like Distributed service attacks, Sybil attacks, vampire attack, defense data drop attacks, decoder attack and soon. So, the development of intrusion detection is important through Artificial intelligence is crucial to prevent the attacks. Mort of exiting model failed to analyze the nature and behavior of feature properties to identify the interrupts due to non-relational communication feature creates more dimension. The data shift problems degrade the poor accuracy in precision and recall rate. To resolve this problem, to propose an advanced Deep scaled Multi perceptron neural network (DSMPNN) classifier based on Enhanced particle swarm intelligence (EPSI) is used for optimal feature selection to improve the classification IDS. Initially the communication logs are collected as to take DarkNet-IDS dataset and preprocessing is carried put to normalize the dataset based on Min-Max normalization. The Traffic intensive defect rate (TIDM) is estimated by observing the feature defect scaling measure. Then the feature section is carried out by EPSI to identify the relation mutual defect feature list and classified with Deep scaled multi perceptron neural network. Finally, the DSMPNN identifies the intrusion classed based on class by reference effectively. The proposed system improves the classification accuracy ass well in higher precision rate, Fl1 rate with higher positive feature limits related to class with redundant false rate compared to the other exiting system.