A Retrospective on Botnet Detection Using Machine Learning Algorithms with Application to Association Rule

  • R. Kiruthika, Dr. K. Selvam
Keywords: Botnet, Features, Detection, network, Association Rule.

Abstract

In new circumstances, detecting a botnet is difficult. A botnet is a massive collection of systems controlled by a bot master using command and control architecture. The bot master's goal is to infect a large number of workstations with the most dangerous poisonous code. Botnet identification utilizing machine learning techniques with association rules is investigated in this article. This analysis came to the conclusion that when feature extraction is done correctly, detection is quite achievable. The dominance of extracted characteristics will result in an extremely high bot detection rate in the network. The paper's conclusion clearly demonstrates that the number of characteristics utilized to detect the bot is unimportant, but feature supremacy is. Using KNN and Naive Bayes, the result achieved after analyzing machine learning algorithms with six different datasets is 99.98 percent. Only 16 characteristics were used to achieve this outcome. The Fuzzy support and confidence employing association rules has an FPR of 0.9, according to the study.

Published
2021-08-16
How to Cite
Dr. K. Selvam, R. K. (2021). A Retrospective on Botnet Detection Using Machine Learning Algorithms with Application to Association Rule . Design Engineering, 8812-8823. Retrieved from http://thedesignengineering.com/index.php/DE/article/view/3435
Section
Articles