Performance Analysis of Deep Neural Network for Intrusion Detection Systems

Intrusion detection system or abbreviated as IDS is an important security system that is used to protect advanced networks used for communication from dangerous threats. These kinds of systems were strategically created to recognize specific rule violations, patterns and signatures. Many great alternatives have been provided by consistent use of machine learning, deep learning algorithms in the subject of network intrusion detection. We can characterize between anomalous and normal behavioral patterns. In this paper, we have done a comparative analysis of our proposed deep learning model with various ML classifiers: Random Forest, Naive Bayesian, Gradient Boosting, Support Vector Machine, Decision Tree, and Logistic Regression. We used Accuracy, Precision and Recall as evaluation metrics for our models. We run our model on various datasets: CICIDS2018, CICIDS2017, UNSW-NB15, NSL-KDD, KDD99 to verify that our model not only identifies particular attacks but performs well on all types of attacks in various datasets. We also draw attention towards a lack of datasets representing the current modern world.