Machine learning for network intrusion detection based on SVM binary classification model

Abstract

Recently, the number of connected machines around the worldwide has become very large, generating a huge amount of data either to be stored or to be communicated. Data protection is a concern for all institutions, it is difficult to manage the masses of data that are susceptible to multiple threats. In this work, we present a novel method of Intrusion Detection System (IDS) based on the detection of anomalies in computer networks. The aim is to use artificial intelligence techniques in the form of Machine Learning (ML) for intrusion detection. For this purpose, we have proposed a Support Vector Machine (SVM) classification model with two kernels, one Polynomial and the other Gaussian. This model is trained and tested with the recent UNSWNB-15 dataset. Regarding the results obtained, we have evaluated our model with six metrics capable of offering all potential threats. As a result, we have achieved a percentage of 94% for the detection rate (DR).