An Artificial Intelligence (AI) Enabled Framework for Cyber Security Using Machine Learning Techniques

Abstract

Cyber security has become very important aspect with respect to security in the contemporary era. The rationale behind this is that, with the emergence of Internet of Things (IoT) use cases, there are millions of connected devices that play crucial role in different applications. Cyber-attacks have been increasing due to the benefits to attackers or adversaries in different means. Therefore, there is need for continuous effort to safeguard cyber space. With respect to different IoT use cases, it is essential to have better solution that is based on machine learning techniques. In this paper an Artificial Intelligence (AI) enabled framework is built for cyber security. The framework is extendible in nature which can support future developments in classifiers. The framework also supports machine learning (ML) models along with feature selection towards cyber security. In other words, it provides support for an AI approach towards safeguarding cyber security. The proposed system is made up of both ML models so as to leverage protection from time to time. It is a generic framework that can be used for any IoT use case provided the inputs from that network of IoT application. The proposed system is made up of both ML models so as to leverage protection from time to time. It is a generic framework that can be used for any IoT use case provided the inputs from that network of IoT application. We proposed an algorithm known as Machine Learning Pipeline for Cyber Attack Detection (MLP-CAD). Experimental results showed that the ML pipeline with underlying techniques could provide better performance. Highest accuracy is achieved by Random Forest with 95.97% accuracy.