Detection of DDoS Attack using Machine Learning Algorithms

Abstract

The utilization of the internet has greatly increased in recent decades, leading to a vulnerability in networking and cybersecurity. One of the most common resulting attacks is Distributed Denial of Service (DDoS), where overwhelming amounts of data are sent to legitimate websites or servers, causing delays or denying access to legitimate users. Single source attacks are known as denial of service (DoS), while attacks from multiple sources, such as a botnet, are considered distributed denial of service (DDoS). In our project, we employed three machine learning algorithms to identify DDoS attacks, and determined the most successful algorithm based on the accuracy metric. We trained and tested our data using the standardized dataset, dataset_sdn, and obtained experimental results. Out of all the algorithms used, the XGBoost algorithm proved to be the most effective with an accuracy of 99.9%. During preprocessing, any missing data was replaced with the column's mean value