A Study on IDS Based CMAC Neuron Network to Improve the Attack Detection Rate

The massive growth of the Internet of Things has brought a lot of attractive benefits because it is going to have a positive impact on life and work through many applications. Besides its advantages, the adoption of massive applications also points the door for attackers to gain cyberattacks on the system. Hence, needed solutions to detect attacks from the edge of the network must be considered to reduce the pressure on the computing elements in core networks. Therefore, approximate approaches to low computational complexity in an Intrusion Detection System (IDS) are being studied to favor limited-resource devices. In this study, a novel IDS based intelligent computation is proposed, the Cerebellar Model Articulation Controller (CMAC) neuron network is chosen to tailor various hardware edge devices. Moreover, to approach edge processing, a feature selection reduction scheme is proposed to reduce the time complexity of the training phase while keeping reasonable accuracy. The experimental results are compared to other previous studies in the same input conditions to high-light the proposed advantages.