Abstract:
In the virtual world, most of the cyber-attacks done by Botnet. The Botnet
is one of the most versatile threats because of its controlling from a remote
place. Most of the existing Botnet detection approaches focused on binary classification
based on traditional machine learning, and these have some limitations. In
this paper, multiclass classification method has been proposed for Botnet detection
based on Artificial Neural Networks with some variations. The proposed model is
used to detect different types of Botnet from a large pool of Botnet families. This
paper has used a dataset consisting of seven different classes to train and test the
model. In this work, we got promising results in terms of accuracy, 99.04%, and other
performance measures. The accuracy of the proposed is better when compared with
other traditional machine learning models when evaluated using the same dataset.