Open AccessIoT botnet detection using machine learning
Author(s) -
M S. B Judyflavia,
P Sowmiyaa,
S Srianvika,
Poojitha Poojitha
Publication year - 2022
Publication title -
international journal of health sciences (ijhs) (en línea)
Language(s) - English
Resource type - Journals
eISSN - 2550-6978
pISSN - 2550-696X
DOI - 10.53730/ijhs.v6ns2.6551
Subject(s) - botnet , computer science , computer security , block (permutation group theory) , layer (electronics) , internet of things , computer network , network security , application layer , network layer , the internet , world wide web , operating system , chemistry , geometry , mathematics , organic chemistry , software deployment
Internet of things (IoT) is a boundless network that connects billions of physical objects implanted with sensors through internet and other various forms of connections. Since there is involvement of many devices, it requires security. To keep the IoT devices safe or making it learn to protect itself with the help of IoT security from the attackers. Common and dangerous malwares detected in IoT devices are botnets, specifically known as IoT botnets that makes one's device fall under malevolent attackers. There are many ways to deal with these attacks and one of the most efficient methodologies used is the multilayer framework, where in the first layer, k-means is used to clear the network traffic and in the second layer, the k-nearest neighbour is applied to block the IP address from entering the network by detecting the C&C server of the botnet. The main drawback is in the second layer, that is, it’s time-consuming and the accuracy is comparatively lower. In this paper, we will use fuzzy c-means for the first layer which will clear the network traffic more efficiently without losing important data. And logistic regression for the second layer, for faster and more accurate detection of botnets.