Rider Optimization based Optimized Deep-CNN towards Attack Detection in IoT
Subramonian Krishna Sarma · 2020
More researchers have made efforts on introducing various models for IoT security frameworks, however, no single IoT framework could assure the optimal security in opposition to varied kinds of attacks. Thereby, this paper intends to propose a new attack detection system by interlinking the Development and Operations (DevOps) concept. The proposed attack detection system involves two stages: Proposed Feature Extraction and Classification. In the first stage, the concatenation of statistical and higher-order statistical features is extracted. These extracted features are then subjected to Deep Convolutional Neural Network (CNN) based classification to identify the attacks in the network. To improve the classification accuracy, the convolution layers in DCNN are optimally tuned using Rider Optimization Algorithm (ROA). Finally, analysis is carried out for validating the betterment of the adopted scheme.