Machine Learning Model for Intrusion Detection

EasyChair Preprint 9337

Abstract

In the field of network security, there is a never-
ending search for cyber-attacks that might disrupt a network.
Furthermore, with the unanticipated emergence and expanded
use of the Internet, hostile network activities are rapidly increas-
ing. It is critical to build a comprehensive intrusion detection
system (IDS) to combat unwanted access to network resources in
order to detect anomalies in the network and secure information.
Intrusion Detection System (IDS) has been an efficient technique
to attain improved security in identifying harmful activity.
Because it is unable to detect all sorts of attacks correctly, current
anomaly detection is frequently linked with high false alarm rates
and only modest accuracy and detection rates. Intrusion detection
systems search for signatures of known attacks or abnormal
activities. Machine learning approaches are taken to approach in
this project by using the KDD-99 Cup and NSL-KDD datasets,
experiment is conducted to evaluate the performance of several
machine learning methods.
Using the NSL-KDD datasets, an experiment is conducted to
analyse the effectiveness of several machine learning methods in
order to design a methodology for creating a Machine Learning
Modal with a higher prediction rate in detecting an attack on
the host network. The results reveal which method worked best
in terms of accuracy, detection rate, and false alarm rate.
The performance of RF, KNN for all attack classes utilising
different feature subsets was above 99 percent. As a result, the
suggested model has a high accuracy rate while also reducing
computational complexity by eliminating unimportant elements

Keyphrases: AirGap Security, IDS, ML, NSL-KDD, Networking

@booklet{EasyChair:9337,
  author    = {Devishree Naidu and Sandesh Sachdev and Vaidik Murarka},
  title     = {Machine Learning Model for Intrusion Detection},
  howpublished = {EasyChair Preprint 9337},
  year      = {EasyChair, 2022}}