A novel weighted support vector machines multiclass classifier based on differential evolution for intrusion detection systems

Abdulla Amin Aburomman, Md. Mamun Ibne Reaz

Research output: Contribution to journalArticle

16 Citations (Scopus)

Abstract

This study compares several methods for creating a multiclass, support vector machines-based (SVM) classifier from a set of binary SVM classifiers. This research aims to identify multiclass SVM models best suited to the intrusion detection task. The methods we compare include one-against-rest SVM (OAR-SVM), one-against-one SVM (OAO-SVM), directed acyclic graph SVM (DAG-SVM), adaptive directed acyclic graph SVM (ADAG-SVM), and error-correcting output code SVM (ECOC-SVM). We also propose a novel approach, based on weighted one-against-rest SVM (WOAR-SVM). Using a set of meta-heuristically generated weights, a WOAR-SVM model is able to compensate for errors in the predictions of individual binary classifiers. In addition, this approach enables seamless integration of several binary hypotheses into a composite, multiclass hypothesis, where each binary classifier may feature a unique set of classification parameters. The results of our experiments on the NSL-KDD benchmark dataset for IDS indicate that WOAR-SVM outperforms the other approaches in terms of overall accuracy.

Original languageEnglish
Pages (from-to)225-246
Number of pages22
JournalInformation Sciences
Volume414
DOIs
Publication statusPublished - 1 Nov 2017

Fingerprint

Intrusion detection
Multi-class
Intrusion Detection
Differential Evolution
Support vector machines
Support Vector Machine
Classifiers
Classifier
Binary
Directed Acyclic Graph
Support vector machine
Differential evolution
Intrusion detection system
Composite
Benchmark
Prediction
Output
Composite materials

Keywords

  • Differential evolution
  • Intrusion detection systems
  • Model selection
  • Multiclass classifiers
  • NSL-KDD
  • Support vector machines

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Theoretical Computer Science
  • Software
  • Computer Science Applications
  • Information Systems and Management
  • Artificial Intelligence

Cite this

A novel weighted support vector machines multiclass classifier based on differential evolution for intrusion detection systems. / Aburomman, Abdulla Amin; Ibne Reaz, Md. Mamun.

In: Information Sciences, Vol. 414, 01.11.2017, p. 225-246.

Research output: Contribution to journalArticle

@article{a9e048c25b374794bd717f38ce7799aa,
title = "A novel weighted support vector machines multiclass classifier based on differential evolution for intrusion detection systems",
abstract = "This study compares several methods for creating a multiclass, support vector machines-based (SVM) classifier from a set of binary SVM classifiers. This research aims to identify multiclass SVM models best suited to the intrusion detection task. The methods we compare include one-against-rest SVM (OAR-SVM), one-against-one SVM (OAO-SVM), directed acyclic graph SVM (DAG-SVM), adaptive directed acyclic graph SVM (ADAG-SVM), and error-correcting output code SVM (ECOC-SVM). We also propose a novel approach, based on weighted one-against-rest SVM (WOAR-SVM). Using a set of meta-heuristically generated weights, a WOAR-SVM model is able to compensate for errors in the predictions of individual binary classifiers. In addition, this approach enables seamless integration of several binary hypotheses into a composite, multiclass hypothesis, where each binary classifier may feature a unique set of classification parameters. The results of our experiments on the NSL-KDD benchmark dataset for IDS indicate that WOAR-SVM outperforms the other approaches in terms of overall accuracy.",
keywords = "Differential evolution, Intrusion detection systems, Model selection, Multiclass classifiers, NSL-KDD, Support vector machines",
author = "Aburomman, {Abdulla Amin} and {Ibne Reaz}, {Md. Mamun}",
year = "2017",
month = "11",
day = "1",
doi = "10.1016/j.ins.2017.06.007",
language = "English",
volume = "414",
pages = "225--246",
journal = "Information Sciences",
issn = "0020-0255",
publisher = "Elsevier Inc.",

}

TY - JOUR

T1 - A novel weighted support vector machines multiclass classifier based on differential evolution for intrusion detection systems

AU - Aburomman, Abdulla Amin

AU - Ibne Reaz, Md. Mamun

PY - 2017/11/1

Y1 - 2017/11/1

N2 - This study compares several methods for creating a multiclass, support vector machines-based (SVM) classifier from a set of binary SVM classifiers. This research aims to identify multiclass SVM models best suited to the intrusion detection task. The methods we compare include one-against-rest SVM (OAR-SVM), one-against-one SVM (OAO-SVM), directed acyclic graph SVM (DAG-SVM), adaptive directed acyclic graph SVM (ADAG-SVM), and error-correcting output code SVM (ECOC-SVM). We also propose a novel approach, based on weighted one-against-rest SVM (WOAR-SVM). Using a set of meta-heuristically generated weights, a WOAR-SVM model is able to compensate for errors in the predictions of individual binary classifiers. In addition, this approach enables seamless integration of several binary hypotheses into a composite, multiclass hypothesis, where each binary classifier may feature a unique set of classification parameters. The results of our experiments on the NSL-KDD benchmark dataset for IDS indicate that WOAR-SVM outperforms the other approaches in terms of overall accuracy.

AB - This study compares several methods for creating a multiclass, support vector machines-based (SVM) classifier from a set of binary SVM classifiers. This research aims to identify multiclass SVM models best suited to the intrusion detection task. The methods we compare include one-against-rest SVM (OAR-SVM), one-against-one SVM (OAO-SVM), directed acyclic graph SVM (DAG-SVM), adaptive directed acyclic graph SVM (ADAG-SVM), and error-correcting output code SVM (ECOC-SVM). We also propose a novel approach, based on weighted one-against-rest SVM (WOAR-SVM). Using a set of meta-heuristically generated weights, a WOAR-SVM model is able to compensate for errors in the predictions of individual binary classifiers. In addition, this approach enables seamless integration of several binary hypotheses into a composite, multiclass hypothesis, where each binary classifier may feature a unique set of classification parameters. The results of our experiments on the NSL-KDD benchmark dataset for IDS indicate that WOAR-SVM outperforms the other approaches in terms of overall accuracy.

KW - Differential evolution

KW - Intrusion detection systems

KW - Model selection

KW - Multiclass classifiers

KW - NSL-KDD

KW - Support vector machines

UR - http://www.scopus.com/inward/record.url?scp=85020708043&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85020708043&partnerID=8YFLogxK

U2 - 10.1016/j.ins.2017.06.007

DO - 10.1016/j.ins.2017.06.007

M3 - Article

VL - 414

SP - 225

EP - 246

JO - Information Sciences

JF - Information Sciences

SN - 0020-0255

ER -