Comparison of Machine Learning Models for Classification of DDoS Attacks

Distributed Denial of Service (DDoS) is one of the major threats for security networks and systems. The DDoS attack is flooding a network with huge packets to weaken the performance of the network. The ability to accurately and classify these attacks is crucial for mitigating the security problem. In this paper, machine learning models are compared in a classification task for detecting benign (normal) and DDoS attack traffic. This paper used 14 different popular classifier models. The models are trained and tested by using ‘DDoS Dataset’ from Kaggle which is the extraction from public IDS Datasets which captured in different years from 2016-2018. The experiment result shown that all the models have good scores in precision, recall, and F1 score with Random Forest algorithm shown the best result at 100% accuracy.