A Study of DDOS Attack Classification Using Machine Learning Classifiers
Soe Kalayar Naing, Tin Thein Thwel
Abstract
Malware threats, security attacks and intrusion are security risks and among them, Distributed Denial of Service (DDOS) attack continues to smash the constructing of integrity and confidentiality of computer systems and network systems. Regardless of the development of security protection techniques, DDOS remains as a severe and challenging issues and hence still need to do more efficient and effective protective methods to expose these DDOS attacks. Specifically, this experiment aims on comparing the classification performance of machine learning algorithms using the open DDOS attack dataset. Some of the outstanding machine learning algorithms, namely, Stochastic Gradient Classifier, Support Vector Machine (SVM), k Nearest Neighbor (kNN), Naïve Bayes and Logistic Regression models are experimented on the open DDOS dataset. According to this experiment, the most outstanding classifier for the classification of DDOS attacks is Logistic Regression classifier. A clear and better understanding of the DDOS dataset for the network traffic is obtained as an additional advantage.