Objectives
■ Analyse the NSL-KDD dataset as a benchmark dataset for intrusion detection
■ Compare three scenarios:
– Multiclass classification (40 labels) – for each different attack
– Grouped classification (5 labels) – attacks are grouped together by kind
– Binary classification (2 labels) – normal and abnormal traffic classification
■ Pre-process the dataset so that it is usable by the models
■ Develop and compare 5 common supervised machine learning classification
algorithms
■ Evaluate results and compare to research
7