A Review of Machine Learning Methods for IoT Network-Centric Anomaly Detection

Anomaly detection in IoT infrastructure is a growing idea in the IoT area. The IoT enables the linking of many devices through the use of wireless and mobile communication technologies. Data received from distributed sensing devices in a certain location is relayed to a central processing center, where it is collected and processed. Data dependability and the quality of IoT services are strongly related. To discover abnormalities, one alternative is to employ Machine Learning (ML) models that have been trained on both normal and aberrant behavior. Only a few are the methods effectively used to discover and evaluate anomalous data with varying degrees of success including Artificial Neural Networks (ANN), Support Vector Machine (SVM), Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN), and the Unsupervised Machine Learning. The findings of the study indicate that, with an accuracy of 99.87%, Unsupervised Machine Learning exceeds the other strategies in terms of performance.