Incremental support vector machine learning: A local approach

Liva Ralaivola; Florence D’alché-Buc

doi:10.1007/3-540-44668-0_46

Incremental support vector machine learning: A local approach

Liva Ralaivola
, Florence D’alché-Buc

LIP6, UPMC Sorbonne Universités - Paris 6

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

In this paper, we propose and study a new on-line algorithm for learning a SVM based on Radial Basis Function Kernel: Local Incremental Learning of SVM or LISVM. Our method exploits the “locality” of RBF kernels to update current machine by only considering a subset of support candidates in the neighbourhood of the input. The determination of this subset is conditioned by the computation of the variation of the error estimate. Implementation is based on the SMO one, introduced and developed by Platt [13]. We study the behaviour of the algorithm during learning when using different generalization error estimates. Experiments on three data sets (batch problems transformed into on-line ones) have been conducted and analyzed.

Original language	English
Title of host publication	Artificial Neural Networks - ICANN 2001 - International Conference, Proceedings
Editors	Kurt Hornik, Georg Dorffner, Horst Bischof
Publisher	Springer Verlag
Pages	322-330
Number of pages	9
ISBN (Print)	3540424865, 9783540446682
DOIs	https://doi.org/10.1007/3-540-44668-0_46
Publication status	Published - 1 Jan 2001
Externally published	Yes
Event	International Conference on Artificial Neural Networks, ICANN 2001 - Vienna, Austria Duration: 21 Aug 2001 → 25 Aug 2001

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	2130
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	International Conference on Artificial Neural Networks, ICANN 2001
Country/Territory	Austria
City	Vienna
Period	21/08/01 → 25/08/01

Access to Document

10.1007/3-540-44668-0_46

Cite this

Ralaivola, L., & D’alché-Buc, F. (2001). Incremental support vector machine learning: A local approach. In K. Hornik, G. Dorffner, & H. Bischof (Eds.), Artificial Neural Networks - ICANN 2001 - International Conference, Proceedings (pp. 322-330). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2130). Springer Verlag. https://doi.org/10.1007/3-540-44668-0_46

Ralaivola, Liva ; D’alché-Buc, Florence. / Incremental support vector machine learning : A local approach. Artificial Neural Networks - ICANN 2001 - International Conference, Proceedings. editor / Kurt Hornik ; Georg Dorffner ; Horst Bischof. Springer Verlag, 2001. pp. 322-330 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{975bdf314a5449ebad9bcd3d357910b0,

title = "Incremental support vector machine learning: A local approach",

abstract = "In this paper, we propose and study a new on-line algorithm for learning a SVM based on Radial Basis Function Kernel: Local Incremental Learning of SVM or LISVM. Our method exploits the “locality” of RBF kernels to update current machine by only considering a subset of support candidates in the neighbourhood of the input. The determination of this subset is conditioned by the computation of the variation of the error estimate. Implementation is based on the SMO one, introduced and developed by Platt [13]. We study the behaviour of the algorithm during learning when using different generalization error estimates. Experiments on three data sets (batch problems transformed into on-line ones) have been conducted and analyzed.",

author = "Liva Ralaivola and Florence D{\textquoteright}alch{\'e}-Buc",

note = "Publisher Copyright: {\textcopyright} Springer-Verlag Berlin Heidelberg 2001.; International Conference on Artificial Neural Networks, ICANN 2001 ; Conference date: 21-08-2001 Through 25-08-2001",

year = "2001",

month = jan,

day = "1",

doi = "10.1007/3-540-44668-0\_46",

language = "English",

isbn = "3540424865",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "322--330",

editor = "Kurt Hornik and Georg Dorffner and Horst Bischof",

booktitle = "Artificial Neural Networks - ICANN 2001 - International Conference, Proceedings",

}

Ralaivola, L & D’alché-Buc, F 2001, Incremental support vector machine learning: A local approach. in K Hornik, G Dorffner & H Bischof (eds), Artificial Neural Networks - ICANN 2001 - International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 2130, Springer Verlag, pp. 322-330, International Conference on Artificial Neural Networks, ICANN 2001, Vienna, Austria, 21/08/01. https://doi.org/10.1007/3-540-44668-0_46

Incremental support vector machine learning: A local approach. / Ralaivola, Liva; D’alché-Buc, Florence.
Artificial Neural Networks - ICANN 2001 - International Conference, Proceedings. ed. / Kurt Hornik; Georg Dorffner; Horst Bischof. Springer Verlag, 2001. p. 322-330 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2130).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Incremental support vector machine learning

T2 - International Conference on Artificial Neural Networks, ICANN 2001

AU - Ralaivola, Liva

AU - D’alché-Buc, Florence

N1 - Publisher Copyright: © Springer-Verlag Berlin Heidelberg 2001.

PY - 2001/1/1

Y1 - 2001/1/1

N2 - In this paper, we propose and study a new on-line algorithm for learning a SVM based on Radial Basis Function Kernel: Local Incremental Learning of SVM or LISVM. Our method exploits the “locality” of RBF kernels to update current machine by only considering a subset of support candidates in the neighbourhood of the input. The determination of this subset is conditioned by the computation of the variation of the error estimate. Implementation is based on the SMO one, introduced and developed by Platt [13]. We study the behaviour of the algorithm during learning when using different generalization error estimates. Experiments on three data sets (batch problems transformed into on-line ones) have been conducted and analyzed.

AB - In this paper, we propose and study a new on-line algorithm for learning a SVM based on Radial Basis Function Kernel: Local Incremental Learning of SVM or LISVM. Our method exploits the “locality” of RBF kernels to update current machine by only considering a subset of support candidates in the neighbourhood of the input. The determination of this subset is conditioned by the computation of the variation of the error estimate. Implementation is based on the SMO one, introduced and developed by Platt [13]. We study the behaviour of the algorithm during learning when using different generalization error estimates. Experiments on three data sets (batch problems transformed into on-line ones) have been conducted and analyzed.

U2 - 10.1007/3-540-44668-0_46

DO - 10.1007/3-540-44668-0_46

M3 - Conference contribution

AN - SCOPUS:84958962423

SN - 3540424865

SN - 9783540446682

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 322

EP - 330

BT - Artificial Neural Networks - ICANN 2001 - International Conference, Proceedings

A2 - Hornik, Kurt

A2 - Dorffner, Georg

A2 - Bischof, Horst

PB - Springer Verlag

Y2 - 21 August 2001 through 25 August 2001

ER -

Ralaivola L, D’alché-Buc F. Incremental support vector machine learning: A local approach. In Hornik K, Dorffner G, Bischof H, editors, Artificial Neural Networks - ICANN 2001 - International Conference, Proceedings. Springer Verlag. 2001. p. 322-330. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/3-540-44668-0_46

Incremental support vector machine learning: A local approach

Abstract

Publication series

Conference

Access to Document

Fingerprint

Cite this