What distinguishes binary from multi-class intrusion detection systems: Observations from experiments

Abstract

Modern world has become prune to technology and security is turning invasive by the day. Thus, capturing personal information or access to remote devices can prove to be horrendous intrusions. This paper focuses on various classification algorithms such as K-nearest neighbor Classifier, Multi Layer Perceptron Classifier, Long Short-Term Memory Classifier and Support Vector Machine Classifiers on the revised KDD cup 99 dataset. Attacks namely DoS (Denial of Service attacks), R2L (Root to Local attacks), U2R (User to Root attack) and Probe (Probing attacks) were monitored. Getting the model ready, we aim to identify the attack types based on the data coming through. The study also showcases Uni-variate, Bi-variate as well as Multivariate analysis on the same. The models were optimized and accuracy was found through measures like F1-score, precision, and recall. Promising results were found.

Keywords

• Network traffic
• Intrusion detection system (IDS)
• Management Information Systems (MIS)
• NSL-KDD dataset