The purpose of this study is to discover anomalies and malicious traffic in
the Internet of Things (IoT) network, which is critical for IoT security, as well as to
keep monitoring and stop undesired traffic flows in the IoT network. For this objective,
a number of researchers have developed several machine learning (ML) approach
models to limit fraudulent traffic flows in the Internet of Things network. On the other
side, due to poor feature selection, some machine learning algorithms are prone to
misclassifying mostly damaging traffic flows. Nonetheless, further study is needed on
the vital problem of how to choose helpful attributes for accurate malicious traffic
identification in the Internet of Things network. As a solution to the problem, an
Artificial Neural Network (ANN) model is proposed. The Area under Curve (AUC)
metric is used to employ the cross-entropy approach to effectively filter features using
the confusion matrix and identify effective features for the chosen Machine Learning
algorithm.
Keywords: Artificial neural network (ANN), IoT Network, Machine learning (ML), Malicious traffic.
