Incremental support vector learning

Analysis, implementation and applications

Pavel Laskov, Christian Gehl, Stefan Krüger, Klaus Muller

Research output: Contribution to journalArticle

239 Citations (Scopus)

Abstract

Incremental Support Vector Machines (SVM) are instrumental in practical applications of online learning. This work focuses on the design and analysis of efficient incremental SVM learning, with the aim of providing a fast, numerically stable and robust implementation. A detailed analysis of convergence and of algorithmic complexity of incremental SVM learning is carried out. Based on this analysis, a new design of storage and numerical operations is proposed, which speeds up the training of an incremental SVM by a factor of 5 to 20. The performance of the new algorithm is demonstrated in two scenarios: learning with limited resources and active learning. Various applications of the algorithm, such as in drug discovery, online monitoring of industrial devices and and surveillance of network traffic, can be foreseen.

Original languageEnglish
Pages (from-to)1909-1936
Number of pages28
JournalJournal of Machine Learning Research
Volume7
Publication statusPublished - 2006 Sep 1
Externally publishedYes

Fingerprint

Support Vector
Support vector machines
Support Vector Machine
Learning systems
Machine Learning
On-line Monitoring
Algorithmic Complexity
Drug Discovery
Online Learning
Active Learning
Network Traffic
Surveillance
Speedup
Scenarios
Resources
Learning
Monitoring
Design

Keywords

  • Drug discovery
  • Incremental SVM
  • Intrusion detection
  • Online learning

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

Incremental support vector learning : Analysis, implementation and applications. / Laskov, Pavel; Gehl, Christian; Krüger, Stefan; Muller, Klaus.

In: Journal of Machine Learning Research, Vol. 7, 01.09.2006, p. 1909-1936.

Research output: Contribution to journalArticle

Laskov, P, Gehl, C, Krüger, S & Muller, K 2006, 'Incremental support vector learning: Analysis, implementation and applications', Journal of Machine Learning Research, vol. 7, pp. 1909-1936.
Laskov, Pavel ; Gehl, Christian ; Krüger, Stefan ; Muller, Klaus. / Incremental support vector learning : Analysis, implementation and applications. In: Journal of Machine Learning Research. 2006 ; Vol. 7. pp. 1909-1936.
@article{4d70b0320958419ca91a76613aac5892,
title = "Incremental support vector learning: Analysis, implementation and applications",
abstract = "Incremental Support Vector Machines (SVM) are instrumental in practical applications of online learning. This work focuses on the design and analysis of efficient incremental SVM learning, with the aim of providing a fast, numerically stable and robust implementation. A detailed analysis of convergence and of algorithmic complexity of incremental SVM learning is carried out. Based on this analysis, a new design of storage and numerical operations is proposed, which speeds up the training of an incremental SVM by a factor of 5 to 20. The performance of the new algorithm is demonstrated in two scenarios: learning with limited resources and active learning. Various applications of the algorithm, such as in drug discovery, online monitoring of industrial devices and and surveillance of network traffic, can be foreseen.",
keywords = "Drug discovery, Incremental SVM, Intrusion detection, Online learning",
author = "Pavel Laskov and Christian Gehl and Stefan Kr{\"u}ger and Klaus Muller",
year = "2006",
month = "9",
day = "1",
language = "English",
volume = "7",
pages = "1909--1936",
journal = "Journal of Machine Learning Research",
issn = "1532-4435",
publisher = "Microtome Publishing",

}

TY - JOUR

T1 - Incremental support vector learning

T2 - Analysis, implementation and applications

AU - Laskov, Pavel

AU - Gehl, Christian

AU - Krüger, Stefan

AU - Muller, Klaus

PY - 2006/9/1

Y1 - 2006/9/1

N2 - Incremental Support Vector Machines (SVM) are instrumental in practical applications of online learning. This work focuses on the design and analysis of efficient incremental SVM learning, with the aim of providing a fast, numerically stable and robust implementation. A detailed analysis of convergence and of algorithmic complexity of incremental SVM learning is carried out. Based on this analysis, a new design of storage and numerical operations is proposed, which speeds up the training of an incremental SVM by a factor of 5 to 20. The performance of the new algorithm is demonstrated in two scenarios: learning with limited resources and active learning. Various applications of the algorithm, such as in drug discovery, online monitoring of industrial devices and and surveillance of network traffic, can be foreseen.

AB - Incremental Support Vector Machines (SVM) are instrumental in practical applications of online learning. This work focuses on the design and analysis of efficient incremental SVM learning, with the aim of providing a fast, numerically stable and robust implementation. A detailed analysis of convergence and of algorithmic complexity of incremental SVM learning is carried out. Based on this analysis, a new design of storage and numerical operations is proposed, which speeds up the training of an incremental SVM by a factor of 5 to 20. The performance of the new algorithm is demonstrated in two scenarios: learning with limited resources and active learning. Various applications of the algorithm, such as in drug discovery, online monitoring of industrial devices and and surveillance of network traffic, can be foreseen.

KW - Drug discovery

KW - Incremental SVM

KW - Intrusion detection

KW - Online learning

UR - http://www.scopus.com/inward/record.url?scp=33745777639&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33745777639&partnerID=8YFLogxK

M3 - Article

VL - 7

SP - 1909

EP - 1936

JO - Journal of Machine Learning Research

JF - Journal of Machine Learning Research

SN - 1532-4435

ER -