Deep CNN-Based IoT Attack Detection with Granger Causality Feature Selection in 6G Networks

Internet of things (IoT) Security is Critical in 6G Networks. Hence, there is a need for advanced security methods that detect cyber attacks in IoT. In this context, this paper proposed a new deep convolutional neural network (CNN) model and Granger Causality Test. We used the Granger Causality Test for the selection of relevant features that affect the detection performance of anomalous traffic patterns that are incorporated in our model. Our proposed model performs well, with an accuracy of 94%. In addition to that, the classification report and confusion matrix provide evidence of the precision and recall of the model, especially for cases in which normal traffic can be distinguished from a variety of attacks.