Quasi-linear support vector machine for nonlinear classification

Bo Zhou, Benhui Chen, Takayuki Furuzuki

Research output: Contribution to journalArticle

16 Citations (Scopus)

Abstract

This paper proposes a so called quasi-linear support vector machine (SVM), which is an SVM with a composite quasi-linear kernel. In the quasi-linear SVM model, the nonlinear separation hyperplane is approximated by multiple local linear models with interpolation. Instead of building multiple local SVM models separately, the quasi-linear SVM realizes the multi local linear model approach in the kernel level. That is, it is built exactly in the same way as a single SVM model, by composing a quasi-linear kernel. A guided partitioning method is proposed to obtain the local partitions for the composition of quasi-linear kernel function. Experiment results on artificial data and benchmark datasets show that the proposed method is effective and improves classification performances.

Original languageEnglish
Pages (from-to)1587-1594
Number of pages8
JournalIEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
VolumeE97-A
Issue number7
DOIs
Publication statusPublished - 2014

Fingerprint

Support vector machines
Support Vector Machine
kernel
Linear Model
Kernel Function
Hyperplane
Linear Function
Partitioning
Interpolation
Interpolate
Composite
Partition
Model
Benchmark
Composite materials
Chemical analysis
Experiment
Experiments

Keywords

  • Interpolation
  • Kernel composition
  • Multiple local linear models
  • Nonlinear separation hyperplane
  • SVM

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Computer Graphics and Computer-Aided Design
  • Applied Mathematics
  • Signal Processing

Cite this

Quasi-linear support vector machine for nonlinear classification. / Zhou, Bo; Chen, Benhui; Furuzuki, Takayuki.

In: IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences, Vol. E97-A, No. 7, 2014, p. 1587-1594.

Research output: Contribution to journalArticle

@article{1354ad4224624ab787b7571ffdbc353c,
title = "Quasi-linear support vector machine for nonlinear classification",
abstract = "This paper proposes a so called quasi-linear support vector machine (SVM), which is an SVM with a composite quasi-linear kernel. In the quasi-linear SVM model, the nonlinear separation hyperplane is approximated by multiple local linear models with interpolation. Instead of building multiple local SVM models separately, the quasi-linear SVM realizes the multi local linear model approach in the kernel level. That is, it is built exactly in the same way as a single SVM model, by composing a quasi-linear kernel. A guided partitioning method is proposed to obtain the local partitions for the composition of quasi-linear kernel function. Experiment results on artificial data and benchmark datasets show that the proposed method is effective and improves classification performances.",
keywords = "Interpolation, Kernel composition, Multiple local linear models, Nonlinear separation hyperplane, SVM",
author = "Bo Zhou and Benhui Chen and Takayuki Furuzuki",
year = "2014",
doi = "10.1587/transfun.E97.A.1587",
language = "English",
volume = "E97-A",
pages = "1587--1594",
journal = "IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences",
issn = "0916-8508",
publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",
number = "7",

}

TY - JOUR

T1 - Quasi-linear support vector machine for nonlinear classification

AU - Zhou, Bo

AU - Chen, Benhui

AU - Furuzuki, Takayuki

PY - 2014

Y1 - 2014

N2 - This paper proposes a so called quasi-linear support vector machine (SVM), which is an SVM with a composite quasi-linear kernel. In the quasi-linear SVM model, the nonlinear separation hyperplane is approximated by multiple local linear models with interpolation. Instead of building multiple local SVM models separately, the quasi-linear SVM realizes the multi local linear model approach in the kernel level. That is, it is built exactly in the same way as a single SVM model, by composing a quasi-linear kernel. A guided partitioning method is proposed to obtain the local partitions for the composition of quasi-linear kernel function. Experiment results on artificial data and benchmark datasets show that the proposed method is effective and improves classification performances.

AB - This paper proposes a so called quasi-linear support vector machine (SVM), which is an SVM with a composite quasi-linear kernel. In the quasi-linear SVM model, the nonlinear separation hyperplane is approximated by multiple local linear models with interpolation. Instead of building multiple local SVM models separately, the quasi-linear SVM realizes the multi local linear model approach in the kernel level. That is, it is built exactly in the same way as a single SVM model, by composing a quasi-linear kernel. A guided partitioning method is proposed to obtain the local partitions for the composition of quasi-linear kernel function. Experiment results on artificial data and benchmark datasets show that the proposed method is effective and improves classification performances.

KW - Interpolation

KW - Kernel composition

KW - Multiple local linear models

KW - Nonlinear separation hyperplane

KW - SVM

UR - http://www.scopus.com/inward/record.url?scp=84903698176&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84903698176&partnerID=8YFLogxK

U2 - 10.1587/transfun.E97.A.1587

DO - 10.1587/transfun.E97.A.1587

M3 - Article

AN - SCOPUS:84903698176

VL - E97-A

SP - 1587

EP - 1594

JO - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

JF - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

SN - 0916-8508

IS - 7

ER -