Machine Learning Techniques to Combat Security Threats in Social Internet of Things

Abstract

A new IoT archetype in which items can create social relationships with one another based on user preferences, resulting in a social platform is known as the Social Internet of Things (SIoT). Machines and gadgets in almost any commercial enterprise may be linked and programmed to offer facts to cloud apps and return end the usage of mobile networks. The act of securing net gadgets and the networks to which they're linked from threats and breaches is referred to as protection inside the Social Internet of Things. Identifying, protecting, and tracking threats, in addition to supporting the restoration of vulnerabilities from some of the technology which could constitute protection hazards, are all methods to offer protection. The blessings of SIoT are apparent, however high-profile assaults, blended with uncertainties regarding approximately protecting good practices and their associated costs, are deterring many companies from imposing it. Here, we are surveying a few machine-learning solutions that address the problem of social internet of things security. Any enterprise trying to secure SIoT devices on a more scalable and efficient basis with automation and aberrant behavior detection can benefit from machine learning. The performance of various machine learning algorithms and tools like Decision Tree (DT), Naive Bayes (NB), K-Nearest Neighbors (KNN), and Artificial Neural Networks (ANN), etc., in discovering the vulnerabilities and threats in SIoT is discussed here.