Intrusion Detection Model for IoT Using Recurrent Kernel Convolutional Neural Network

In communication and information technology, the Internet of Things (IoT) creates an enormous amount of data traffic that permits data analysis to expose and detect unusual network load and hidden trends. There are different existing DL (Deep Learning) classification methods applied for IDM (Intrusion Detection Model) detection, but it has some limitations, such as being difficult for security and more complex. In order to overcome these problems, the proposed work used the DL based RK.CNN-MMBO (Recurrent Kernel Convolution Neural Network-Modified Monarch Butterfly Optimization) model to identify the attacks in the network. In the pre-processing stage, the data are pre-processed by min_max normalization. After pre-processing, the optimal features are selected by the IBRO (Improved Battle Royale Optimization) algorithm. Then the selected features are classified using the DL classifier RKCNN (Recurrent kernel convolutional neural network). Moreover, MMBO (Modified Monarch Butterfly Optimization) improves the classifier's performance in attack detection. The proposed work used two different types of datasets for experimental validation, N-BaIoT and CICIDS-2017. The proposed IDM classification using the N-BaIoT dataset with the RKCNN-MMBO model attains the accuracy of (99.96%), and the CICIDS-2017 dataset with the RKCNN-MMBO model obtains the accuracy value of (99.95%). The proposed RKCNN-MMBO classifier model obtained higher accuracy in the detection of IDM than other methods.