Regression algorithms for efficient detection and prediction of DDoS attacks

Abstract

In the ICT era the need of depth investigation and analysis is required on network traffic. The analysis should focus on detecting DDoS attacks. In the 21st century the use of communication or transactions are completely doing through online, the political activists, and international cyber terrorists are choosing the DDoS attacks as a powerful weapon for their illegal an un ethical activities. It is impossible to the human being to identify all these unethical activities, hence the need of machine based algorithms are required. In this paper we used GLM, GBM, NN, RF regression algorithms for detection and prediction of DDoS attacks, and also proved that by using regression algorithms we observed more accurate result than using KNN SVM algorithm.