IDS in Telecommunication Network Using PCA

TL;DR

This paper proposes two anomaly detection intrusion detection systems using PCA, achieving high success and detection rates on the NSL-KDD dataset to improve network security.

Contribution

It introduces two PCA-based IDS models, one for network and one for host intrusion detection, with demonstrated high effectiveness on benchmark data.

Findings

Network IDS success rate: 0.9161

Host IDS success rate: 0.8493

Detection rates: 0.9288 (network), 0.9628 (host)

Abstract

Data Security has become a very serious part of any organizational information system. Internet threats have become more intelligent so it can deceive the basic security solutions such as firewalls and antivirus scanners. To enhance the overall security of the network an additional security layer such as intrusion detection system (IDS) has to be added. The anomaly detection IDS is a type of IDS that can differentiate between normal and abnormal in the data monitored. This paper proposes two types of IDS, one of them can be used as a network intrusion detection system (NIDS) with overall success (0.9161) and high detection rate (0.9288) and the other type can also be used as a host intrusion detection system (HIDS) with overall success (0.8493) and very high detection rate (0.9628) using NSL-KDD data set.