ML-IBotD: Machine Learning based Intelligent Botnet Detection

Abstract

With the advancements in communication technologies, an abundance of smart devices and internet-based applications in every walk of human life has resulted in the production of a huge number of data transmissions over the internet. In line with this emergence, the number of cybersecurity attacks is also rising. Among notable network attacks like mal ware, phishing, etc., we focused on botnet attacks which can cause huge damage on a large scale because botnet works in network form which appears as an adverse risk for the internet. In the botnet, there are many compromised systems known as bots controlled by the botmaster. On the other hand, Machine Learning (ML) is playing an important role in the detection of such network attacks with notable accuracy. In this paper, we select a dataset of CIC-IDS2017 due to its real interpretation of botnets. Then flows are extracted and then relevant four features are selected from the flows. In this paper, we apply four classifiers of SVM, KNN, DT, and Ensemble classifier on a real dataset of CIC-IDS2017. The highest achieved testing accuracy is 99.56% with the Ensemble classifier.