The Detection of Mirai Botnet Attack on the Internet of Things (IoT) Device Using Support Vector Machine (SVM) Model
Revaldi Gilang Azhari, Vera Suryani, Rizka Reza Pahlevi, Aulia Arif Wardana
Abstract
There are various types of attacks on IoT devices, one of which is Mirai botnet that attacks a number of IoT devices such as web cameras, security cameras, and routers. Even it has caused internet downtime for the users in Europe and North America. Hence, it is deemed necessary to have a system to detect Mirai botnet attack using the Support Vector Machine (SVM) model. This model becomes a solution to detect Mirai botnet attacks for having good generalization performance when the parameter is properly configured in modeling the training set (training dataset) and getting good classification results without a lot of training data. The results of the analysis from the use of the SVM model showed the accuracy of 92.91% with a linear kernel and max_iter 10000. Also, the performance of the Support Vector Machine (SVM) model based on the precision parameter obtained a value of 94.13% with max_iter 10000 and the recall parameter resulted in a value of 91.10% with max_iter 10000.