A novel SVM-kNN-PSO ensemble method for intrusion detection system

Abdulla Amin Aburomman, Md. Mamun Ibne Reaz

Research output: Contribution to journalArticle

134 Citations (Scopus)

Abstract

In machine learning, a combination of classifiers, known as an ensemble classifier, often outperforms individual ones. While many ensemble approaches exist, it remains, however, a difficult task to find a suitable ensemble configuration for a particular dataset. This paper proposes a novel ensemble construction method that uses PSO generated weights to create ensemble of classifiers with better accuracy for intrusion detection. Local unimodal sampling (LUS) method is used as a meta-optimizer to find better behavioral parameters for PSO. For our empirical study, we took five random subsets from the well-known KDD99 dataset. Ensemble classifiers are created using the new approaches as well as the weighted majority algorithm (WMA) approach. Our experimental results suggest that the new approach can generate ensembles that outperform WMA in terms of classification accuracy.

Original languageEnglish
Pages (from-to)360-372
Number of pages13
JournalApplied Soft Computing Journal
Volume38
DOIs
Publication statusPublished - 1 Jan 2016

Fingerprint

Intrusion detection
Particle swarm optimization (PSO)
Classifiers
Learning systems
Sampling

Keywords

  • Ensemble
  • k-NN
  • LUS
  • PSO
  • SVM
  • Weighted majority voting (WMV)

ASJC Scopus subject areas

  • Software

Cite this

A novel SVM-kNN-PSO ensemble method for intrusion detection system. / Aburomman, Abdulla Amin; Ibne Reaz, Md. Mamun.

In: Applied Soft Computing Journal, Vol. 38, 01.01.2016, p. 360-372.

Research output: Contribution to journalArticle

@article{0ac1c9320829499390478cd6503d0277,
title = "A novel SVM-kNN-PSO ensemble method for intrusion detection system",
abstract = "In machine learning, a combination of classifiers, known as an ensemble classifier, often outperforms individual ones. While many ensemble approaches exist, it remains, however, a difficult task to find a suitable ensemble configuration for a particular dataset. This paper proposes a novel ensemble construction method that uses PSO generated weights to create ensemble of classifiers with better accuracy for intrusion detection. Local unimodal sampling (LUS) method is used as a meta-optimizer to find better behavioral parameters for PSO. For our empirical study, we took five random subsets from the well-known KDD99 dataset. Ensemble classifiers are created using the new approaches as well as the weighted majority algorithm (WMA) approach. Our experimental results suggest that the new approach can generate ensembles that outperform WMA in terms of classification accuracy.",
keywords = "Ensemble, k-NN, LUS, PSO, SVM, Weighted majority voting (WMV)",
author = "Aburomman, {Abdulla Amin} and {Ibne Reaz}, {Md. Mamun}",
year = "2016",
month = "1",
day = "1",
doi = "10.1016/j.asoc.2015.10.011",
language = "English",
volume = "38",
pages = "360--372",
journal = "Applied Soft Computing",
issn = "1568-4946",
publisher = "Elsevier BV",

}

TY - JOUR

T1 - A novel SVM-kNN-PSO ensemble method for intrusion detection system

AU - Aburomman, Abdulla Amin

AU - Ibne Reaz, Md. Mamun

PY - 2016/1/1

Y1 - 2016/1/1

N2 - In machine learning, a combination of classifiers, known as an ensemble classifier, often outperforms individual ones. While many ensemble approaches exist, it remains, however, a difficult task to find a suitable ensemble configuration for a particular dataset. This paper proposes a novel ensemble construction method that uses PSO generated weights to create ensemble of classifiers with better accuracy for intrusion detection. Local unimodal sampling (LUS) method is used as a meta-optimizer to find better behavioral parameters for PSO. For our empirical study, we took five random subsets from the well-known KDD99 dataset. Ensemble classifiers are created using the new approaches as well as the weighted majority algorithm (WMA) approach. Our experimental results suggest that the new approach can generate ensembles that outperform WMA in terms of classification accuracy.

AB - In machine learning, a combination of classifiers, known as an ensemble classifier, often outperforms individual ones. While many ensemble approaches exist, it remains, however, a difficult task to find a suitable ensemble configuration for a particular dataset. This paper proposes a novel ensemble construction method that uses PSO generated weights to create ensemble of classifiers with better accuracy for intrusion detection. Local unimodal sampling (LUS) method is used as a meta-optimizer to find better behavioral parameters for PSO. For our empirical study, we took five random subsets from the well-known KDD99 dataset. Ensemble classifiers are created using the new approaches as well as the weighted majority algorithm (WMA) approach. Our experimental results suggest that the new approach can generate ensembles that outperform WMA in terms of classification accuracy.

KW - Ensemble

KW - k-NN

KW - LUS

KW - PSO

KW - SVM

KW - Weighted majority voting (WMV)

UR - http://www.scopus.com/inward/record.url?scp=84945936408&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84945936408&partnerID=8YFLogxK

U2 - 10.1016/j.asoc.2015.10.011

DO - 10.1016/j.asoc.2015.10.011

M3 - Article

AN - SCOPUS:84945936408

VL - 38

SP - 360

EP - 372

JO - Applied Soft Computing

JF - Applied Soft Computing

SN - 1568-4946

ER -