Extending Network-based Intrusion Detection Systems Through Non- Intrusive Machine Learning-Based Approach

Abstract

this study provides an alternative approach in training and testing datasets with machine learning and studies the effectiveness of the performance. This includes an understanding of detection rate, accuracy rate, and feature selection. Therefore, this study is used to survey the factors relevant to the development of a viable IDS, one compatible with machine learning techniques and compare that result to an actual open source network- based intrusion detection system. This is done in order to establish that a Non-intrusive IDS can be as effective as network-based Intrusion detection.