Dimensionality reduction for denial of service detection problems using RBFNN output sensitivity

W. W Y Ng; Kow Chuen Chang; Daniel S. Yeung

Dimensionality reduction for denial of service detection problems using RBFNN output sensitivity

W. W Y Ng, Kow Chuen Chang, Daniel S. Yeung

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

27 Citations (Scopus)

Abstract

In this paper, we have presented a feature importance ranking methodology based on the stochastic radial basis function neural network output sensitivity measure and have shown, for the 10% training set of the DARPA network intrusion detection data set prepared by MIT Lincoln Labs, that 33 out of 41 features (more than 80% dimensionality reduction) can be removed without causing great harm to the classification accuracy of denial of service (DoS) attacks and normal packets (false positives rise from 0.7% to 0.93%). The reduced feature subset leads to more generalized and less complex model for classifying DoS and Normal. Exploratory discussions on the relevancy of the selected features and the DoS attack types are presented.

Original language	English
Title of host publication	International Conference on Machine Learning and Cybernetics
Pages	1293-1298
Number of pages	6
Volume	2
Publication status	Published - 1 Dec 2003
Event	International Conference on Machine Learning and Cybernetics - Xi'an, China Duration: 2 Nov 2003 → 5 Nov 2003

Conference

Conference	International Conference on Machine Learning and Cybernetics
Country/Territory	China
City	Xi'an
Period	2/11/03 → 5/11/03

Keywords

Denial of Service (DoS)
Feature Selection
Network Intrusion Detection
RBFNN
Sensitivity Analysis

ASJC Scopus subject areas

Artificial Intelligence

Cite this

@inproceedings{c0d573620071422c8d438707b45e8067,

title = "Dimensionality reduction for denial of service detection problems using RBFNN output sensitivity",

abstract = "In this paper, we have presented a feature importance ranking methodology based on the stochastic radial basis function neural network output sensitivity measure and have shown, for the 10% training set of the DARPA network intrusion detection data set prepared by MIT Lincoln Labs, that 33 out of 41 features (more than 80% dimensionality reduction) can be removed without causing great harm to the classification accuracy of denial of service (DoS) attacks and normal packets (false positives rise from 0.7% to 0.93%). The reduced feature subset leads to more generalized and less complex model for classifying DoS and Normal. Exploratory discussions on the relevancy of the selected features and the DoS attack types are presented.",

keywords = "Denial of Service (DoS), Feature Selection, Network Intrusion Detection, RBFNN, Sensitivity Analysis",

author = "Ng, {W. W Y} and Chang, {Kow Chuen} and Yeung, {Daniel S.}",

year = "2003",

month = dec,

day = "1",

language = "English",

isbn = "0780378652",

volume = "2",

pages = "1293--1298",

booktitle = "International Conference on Machine Learning and Cybernetics",

note = "International Conference on Machine Learning and Cybernetics ; Conference date: 02-11-2003 Through 05-11-2003",

}

Dimensionality reduction for denial of service detection problems using RBFNN output sensitivity. / Ng, W. W Y; Chang, Kow Chuen; Yeung, Daniel S.
International Conference on Machine Learning and Cybernetics. Vol. 2 2003. p. 1293-1298.

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

TY - GEN

T1 - Dimensionality reduction for denial of service detection problems using RBFNN output sensitivity

AU - Ng, W. W Y

AU - Chang, Kow Chuen

AU - Yeung, Daniel S.

PY - 2003/12/1

Y1 - 2003/12/1

N2 - In this paper, we have presented a feature importance ranking methodology based on the stochastic radial basis function neural network output sensitivity measure and have shown, for the 10% training set of the DARPA network intrusion detection data set prepared by MIT Lincoln Labs, that 33 out of 41 features (more than 80% dimensionality reduction) can be removed without causing great harm to the classification accuracy of denial of service (DoS) attacks and normal packets (false positives rise from 0.7% to 0.93%). The reduced feature subset leads to more generalized and less complex model for classifying DoS and Normal. Exploratory discussions on the relevancy of the selected features and the DoS attack types are presented.

AB - In this paper, we have presented a feature importance ranking methodology based on the stochastic radial basis function neural network output sensitivity measure and have shown, for the 10% training set of the DARPA network intrusion detection data set prepared by MIT Lincoln Labs, that 33 out of 41 features (more than 80% dimensionality reduction) can be removed without causing great harm to the classification accuracy of denial of service (DoS) attacks and normal packets (false positives rise from 0.7% to 0.93%). The reduced feature subset leads to more generalized and less complex model for classifying DoS and Normal. Exploratory discussions on the relevancy of the selected features and the DoS attack types are presented.

KW - Denial of Service (DoS)

KW - Feature Selection

KW - Network Intrusion Detection

KW - RBFNN

KW - Sensitivity Analysis

UR - http://www.scopus.com/inward/record.url?scp=1542316134&partnerID=8YFLogxK

M3 - Conference article published in proceeding or book

SN - 0780378652

VL - 2

SP - 1293

EP - 1298

BT - International Conference on Machine Learning and Cybernetics

T2 - International Conference on Machine Learning and Cybernetics

Y2 - 2 November 2003 through 5 November 2003

ER -

Dimensionality reduction for denial of service detection problems using RBFNN output sensitivity

Abstract

Conference

Keywords

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this