Malicious Traffic Flow Detection in IOT Using Ml Based Algorithms

Abstract

Identifying the malicious traffic flows in the Internet of Things (IoT) is very important to monitor and avoid unwanted errors or unwanted flows in the network. So, for security in this network, various machine learning algorithms (ML) have been introduced by various analysts to avoid this flow of errors in the network. However, owing to the unsuitable selection of features, the ML models introduced previously suffer from misclassification errors. Therefore, there arises a need to study the problem of feature selection in more depth to accurately predict traffic flow observations in the network. To overcome this problem, a new structure in machine learning (ML) is introduced. Thus, a novel feature selection metric called CorrAUC is suggested. Based on this metric approach, a new feature selection algorithm, CorrAUC, is developed and designed. It is based on a wrapper technique to accurately filter features to predict the flow of traffic. Then, we applied a multicriteria decision method called VIKOR to validate the features selected for recognizing the flow of traffic errors in the network. We evaluate our approach using the NSL-KDD dataset and three different ML algorithms.