Improving Intrusion Detection System By Utilizing Stacking Based Convolutional Neural Network Ensemble Classifier
Pavan Kumar S, A Shriyans, Ashwath Nandan Rajkumar, Srinath Hariharan, S Kanthimathi · 2024
This research work proposes a deep learning-based ensemble classifier which is utilized for network intrusion detection. This system solves the problem of network attacks by preventing the users from attackers by classifying the packets into safe and different categories of attacks. The proposed framework utilizes the predictions of the individual classifiers and learn from them to make its own predictions. The individual classifiers involve different variations of convolutional neural networks which produces feature maps by learning the patterns present in the data. The proposed methodology is evaluated in terms of accuracy, precision, recall and f1-score and the values are obtained as 98.22%, 97.32%, 98.22% and 97.77% respectively which indicates that the proposed methodology is effective in preventing the system from intruders.