Least Square Support Vector Machine based Intrusion Detection System in IoT
A. Parveen Akhther, A. Maryposonia, V. S. Prasanth
Abstract
The IoT (Internet of Things) is an ever-expanding system of interconnected computing devices. It’s a term for when real-world items can communicate with one another via data transfers. The availability and dependability of its activities are crucial to the day-to-day functioning of modern society. As a result, it’s important to find a solution to the problem of ensuring secured data transmission across Internet of Things. An Intrusion Detection System (IDS) is a piece of software designed to keep an eye on a network or computer system and to address any suspicious activity. New threats and countermeasures must be incorporated into the network’s protected evolution. The primary function of IDS is to protect valuable resources against intrusion. Intruders can be spotted using various intrusion detection techniques, strategies, and algorithms. This major purpose of this study is to propose a comprehensive system for identifying intrusion risks by applying Least Squares Support Vector Machine approach. After receiving input, the following step is preprocessing, which was accomplished via Normalization, Discretization, and Feature selection in this study. These techniques are used for data preprocessing and feature selection. Training the model using LSSVM is the final stage. As compared to other models like SVM and RF, the proposed model attains an accuracy of roughly 97.7%.