Scribd
Upload
39 views1 page

Literature Survey

This document summarizes several papers on network intrusion detection systems. The papers propose different algorithms and techniques like hybrid approaches using frequency episode extraction and chi-square analysis, gradient boosting and clustering, artificial neural networks and support vector machines, correlation-based feature selection and ensemble learning. The papers evaluate the proposed models on datasets like NSL-KDD, achieving high accuracy and detection rates in most cases. However, some models showed limitations like slow detection rates or inability to address zero-day attacks.

Uploaded by

Learn Samadesh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
39 views1 page

Literature Survey

This document summarizes several papers on network intrusion detection systems. The papers propose different algorithms and techniques like hybrid approaches using frequency episode extraction and chi-square analysis, gradient boosting and clustering, artificial neural networks and support vector machines, correlation-based feature selection and ensemble learning. The papers evaluate the proposed models on datasets like NSL-KDD, achieving high accuracy and detection rates in most cases. However, some models showed limitations like slow detection rates or inability to address zero-day attacks.

Uploaded by

Learn Samadesh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 1

TITLE AUTHOR OBJECTIVES Software / ALGORITHM / DATASET / RESULT (ACCURACY or ANY FINDING / Achievement DRAWBACK

Hardware METHODOLOGY / Source of OTHER PARAMETER)


Requirements / Techniques / Methods Input
Programming
Real Time Hybrid I. Dutt et al 1. Hybrid approach that KDD Cup’99 True Positive (TP)=92.65% The hybrid approach used The model showed slow
Intrusion Detection (2018) 2. comprise : helped in achieving a high detection rate when it was
System[11] 1)Frequency Episode Classification Rate = 80% detection rate applied on a big size data
Extraction Accuracy = 75%
2) Chi-Square Analysis
Network Intrusion Verma et al 1) Extreme Gradient Boosting NSL-KDD XGBoost with Clustering=84.253% The work showed that anomaly --
Detection using (2018) (XGBoost) XGBoost without detection has a room in
Clustering and 2) Adaptive Boosting Clustering=80.238 improving its false positive
Gradient boosting[12] (AdaBoost) AdaBoost with
Clustering=82.011%
AdaBoost without
Clustering=80.731%
Network Intrusion Kazi Abu 1) Artificial Neural Network NSL-KDD ANN=94.02% High accuracy was achieved due Inability of the work to
Detection using Taher et al (ANN) to the application of feature address the issue of zero
Supervised Machine (2019) 2) Support Vector Machine selection day attack due to high
Learning Technique (SVM) false positive rate
with feature
selection[13]
Building an Efficient Zhou et al 1)Correlation based feature 1)NSL-KDD Ensemble (NSL-KDD)=99.80% The model was evaluated on False positive was
Intrusion Detection (2019) selection (CFS-BA) 2)AWID Ensemble (AWID)=99.50% three different datasets and observed in CIC-
System[14] 2)Ensemble approach that 3)CIC- Ensemble (CIC-IDS2017)=99.90% returns with an improved IDS2017 dataset
comprise: C4.5, Random IDS2017 efficiency and high detection
Forest (RF) and Forest by rate.
Penalizing (Forest PA)

Anomaly Based Vinoth and 1)Decision Tree NSL-KDD Ensemble=85.20% The work handled imbalanced Trial on the most updated
Network Intrusion Kamatchi 2) Bayes Classifier data and selected only required datasets needs to be
Detection using (2020) 3) RNN-LSTM features which greatly helped in carried out
Ensemble Machine 4)Random Forest reducing high false positive rate
Learning Technique[16] 5) Ensemble of the 4
classifiers
Network Intrusion Bhavani et al 1) Random Forest (RF) NSL-KDD RF=95.323% Easily implemented Slow detection rate and
Detection System (2020) 2) Decision Tree (DT) DT=81.868% high false positive
using Random Forest
and Decision Tree
Machine Learning
Techniques[17]

LITERATURE SURVEY

You might also like