Available under a Creative Commons Attribution Non-Commercial Share Alike 4.0 International Licence
Botnets pose a significant and growing risk to modern networks. Detection of botnets remains an important area of open research in order to prevent the proliferation of botnets and to mitigate the damage that can be caused by botnets that have already been established. Botnet detection can be broadly categorised into two main categories: signature-based detection and anomaly-based detection. This paper sets out to measure the accuracy, false-positive rate, and false-negative rate of four algorithms that are available in Weka for anomaly-based detection of a dataset of HTTP and IRC botnet data. The algorithms that were selected to detect botnets in the Weka environment are J48, naïve Bayes, random forest, and UltraBoost. The dataset was generated using a realistic network environment by The University of New South Wales, Canberra. The findings showed that botnet behaviours from the selected dataset could be detected by Weka with a high degree of accuracy and low false-positive rate. With all features included, the random forest algorithm was found to achieve the highest accuracy with 96.70%, and the algorithm that attained the lowest false-positive rates was also random forest with 0.008. With a reduced feature set of IP addresses and ports, the random forest algorithm attained the highest accuracy and precision and lowest false-positive rate. With only information regarding packets per second being sent and received, J48 was this time the most accurate with its predictions and attained the highest precision.
Porteous, C. (2022). An Analysis on Network Flow-Based IoT Botnet Detection Using Weka. Technological University Dublin. DOI: 10.21427/NW1R-7R89