Intrusion Detection System in IoT Network by using Metaheuristic Algorithm with Machine Learning Dimensional Reduction Technique

The network security is a main problem in any disseminated framework. To providing the safe and secure network we have been proposed anomaly detection system against from the suspicious attacks. The essential goal of this exploration is to plan effective IDS for IoT organization. The interruption discovery assumes a fundamental part in recognizing various assaults on IoT and upgrades the IoT performance. In this paper, one of the metaheuristic algorithms GSO (glow-worm swarm optimization) algorithm add with machine learning technique PCA (principal component analysis) used for providing the IDS. Therefore, the metaheuristic-based intrusion discovery model to recognize malicious activities by utilizing the NSL-KDD dataset. Dimensional reduction technique Principal component analysis is utilized for extraction of features and GSO algorithm is utilized to classify the various class of malicious attacks in the dataset. We have considered some metrics like precision, detection rate, recall, accuracy and FAR are computed. The suggested method evaluated better in each parameter compared with other existing techniques.

I. Introduction

The RFID (Radio Frequency Identification Development) group first proposed the Internet of Things (IoT) in 1999, and it has since emerged as crucial to the real world's advancement of phones, implanted and all-encompassing communication, distributed computing, and information research [1]. The IoT is predicted to play a main role in bringing all needs of our daily lives together. It includes a variety of areas such as vehicles, health care, sports, entertainment, and upscale houses. The widespread use of IoT improves some daily activities, expands social interactions with people and objects, and advances how people interact with the environment [2].