Design an Intrusion Detection System based on feature selection using ML Algorithm

Abstract

With the development of the Internet and Technology, cyber-attacks are growing rapidly thus cyber security is an important aspect which needs to be given attention. There are several types of attacks occurring across the internet like: Denial of Service (DoS) , R2L (Root to Local attacks), U2R (User to Root attack), Probe (Probing attacks) , DNS Spoofing attack etc for which ways to identify the vulnerabilities have to be found. Machine learning, an emerging field in the current time has many techniques that can be adopted in various domains where it shows its dominance over many traditional algorithms. These techniques can be exploited in the field of cyber security with the aim of detecting intrusions which can support or even replace the first level of security analysts. This paper aims in building a model for intrusion detection by applying various machine learning algorithms on the selected features obtained through the modelling process. NSL-KDD dataset is used in order to build IDS. The paper provides a brief description of various machine learning algorithms like SVM, Random Forests, Decision Tree in order to study the of attacks on the prescribed dataset using the selected critical features which aims in improving the accuracy. The paper also provides a comparative study of the algorithms used to determine the algorithm that works the best.