DDoS Attack Detection Analysis Using Ensemble Learning with XGBoost and AdaBoost Algorithms

Cyber attacks have been growing rapidly in every area of human life. A security system is necessary to prevent cyber attacks from causing chaos in the networks. DDOS is a well-known cyber attack that may intrude the networks. These attacks may leak sensitive data or disrupt operational performance causing enormous financial loss to the victim. The ensemble model is an important tool to enhance the learning process of machine learning models. This model will combine XGBoost and AdaBoost algorithms using XGBoost Classifier, AdaBoost Classifier, Decision Tree Classifier, and Voting Classifier. XGBoost and AdaBoost algorithms are used to analyze the data test, which will then be compared with the ensemble model. The best outcomes from the ensemble model yielded 94.88% accuracy, the XGBoost algorithm yielded 92.92% accuracy and the AdaBoost algorithm yielded 92.96% accuracy. An ensemble model produces an enhanced significant accuracy around 2.03% - 2.06% concluding to the experiment results.

I. Introduction

Cyberattack has grown significantly with the widespread use of the IoT (Internet of Things). A vast network of devices known as the "Internet of Things" can communicate with one another and the internet to exchange data and services without the assistance of humans[1]. The most frequent assaults, according to IBM 2022 [2], are server access, DDoS, RAT, insiders, and credential harvesting activities.