A kernel level composition of multiple local classifiers for nonlinear classification

Weite Li, Bo Zhou, Takayuki Furuzuki

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Kernel functions based machine learning algorithms have been extensively studied over the past decades with successful applications in a variety of real-world tasks. In this paper, we formulate a kernel level composition method to embed multiple local classifiers (kernels) into one kernel function, so as to obtain a more flexible data-dependent kernel. Since such composite kernels are composed by multiple local classifiers interpolated with several localizing gating functions, a specific learning process is also introduced in this paper to pre-determine their parameters. Experimental results are provided to validate two major perspectives of this paper. Firstly, the introduced learning process is effective to detect local information, which is essential for the parameter pre-determination of the localizing gating functions. Secondly, the proposed composite kernel has a capacity to improve classification performance.

Original languageEnglish
Title of host publication2016 International Joint Conference on Neural Networks, IJCNN 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages3845-3850
Number of pages6
Volume2016-October
ISBN (Electronic)9781509006199
DOIs
Publication statusPublished - 2016 Oct 31
Event2016 International Joint Conference on Neural Networks, IJCNN 2016 - Vancouver, Canada
Duration: 2016 Jul 242016 Jul 29

Other

Other2016 International Joint Conference on Neural Networks, IJCNN 2016
CountryCanada
CityVancouver
Period16/7/2416/7/29

Fingerprint

Classifiers
Chemical analysis
Composite materials
Learning algorithms
Learning systems

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence

Cite this

Li, W., Zhou, B., & Furuzuki, T. (2016). A kernel level composition of multiple local classifiers for nonlinear classification. In 2016 International Joint Conference on Neural Networks, IJCNN 2016 (Vol. 2016-October, pp. 3845-3850). [7727696] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IJCNN.2016.7727696

A kernel level composition of multiple local classifiers for nonlinear classification. / Li, Weite; Zhou, Bo; Furuzuki, Takayuki.

2016 International Joint Conference on Neural Networks, IJCNN 2016. Vol. 2016-October Institute of Electrical and Electronics Engineers Inc., 2016. p. 3845-3850 7727696.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Li, W, Zhou, B & Furuzuki, T 2016, A kernel level composition of multiple local classifiers for nonlinear classification. in 2016 International Joint Conference on Neural Networks, IJCNN 2016. vol. 2016-October, 7727696, Institute of Electrical and Electronics Engineers Inc., pp. 3845-3850, 2016 International Joint Conference on Neural Networks, IJCNN 2016, Vancouver, Canada, 16/7/24. https://doi.org/10.1109/IJCNN.2016.7727696
Li W, Zhou B, Furuzuki T. A kernel level composition of multiple local classifiers for nonlinear classification. In 2016 International Joint Conference on Neural Networks, IJCNN 2016. Vol. 2016-October. Institute of Electrical and Electronics Engineers Inc. 2016. p. 3845-3850. 7727696 https://doi.org/10.1109/IJCNN.2016.7727696
Li, Weite ; Zhou, Bo ; Furuzuki, Takayuki. / A kernel level composition of multiple local classifiers for nonlinear classification. 2016 International Joint Conference on Neural Networks, IJCNN 2016. Vol. 2016-October Institute of Electrical and Electronics Engineers Inc., 2016. pp. 3845-3850
@inproceedings{3713c2dbe17644229b49fe79071d2086,
title = "A kernel level composition of multiple local classifiers for nonlinear classification",
abstract = "Kernel functions based machine learning algorithms have been extensively studied over the past decades with successful applications in a variety of real-world tasks. In this paper, we formulate a kernel level composition method to embed multiple local classifiers (kernels) into one kernel function, so as to obtain a more flexible data-dependent kernel. Since such composite kernels are composed by multiple local classifiers interpolated with several localizing gating functions, a specific learning process is also introduced in this paper to pre-determine their parameters. Experimental results are provided to validate two major perspectives of this paper. Firstly, the introduced learning process is effective to detect local information, which is essential for the parameter pre-determination of the localizing gating functions. Secondly, the proposed composite kernel has a capacity to improve classification performance.",
author = "Weite Li and Bo Zhou and Takayuki Furuzuki",
year = "2016",
month = "10",
day = "31",
doi = "10.1109/IJCNN.2016.7727696",
language = "English",
volume = "2016-October",
pages = "3845--3850",
booktitle = "2016 International Joint Conference on Neural Networks, IJCNN 2016",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

TY - GEN

T1 - A kernel level composition of multiple local classifiers for nonlinear classification

AU - Li, Weite

AU - Zhou, Bo

AU - Furuzuki, Takayuki

PY - 2016/10/31

Y1 - 2016/10/31

N2 - Kernel functions based machine learning algorithms have been extensively studied over the past decades with successful applications in a variety of real-world tasks. In this paper, we formulate a kernel level composition method to embed multiple local classifiers (kernels) into one kernel function, so as to obtain a more flexible data-dependent kernel. Since such composite kernels are composed by multiple local classifiers interpolated with several localizing gating functions, a specific learning process is also introduced in this paper to pre-determine their parameters. Experimental results are provided to validate two major perspectives of this paper. Firstly, the introduced learning process is effective to detect local information, which is essential for the parameter pre-determination of the localizing gating functions. Secondly, the proposed composite kernel has a capacity to improve classification performance.

AB - Kernel functions based machine learning algorithms have been extensively studied over the past decades with successful applications in a variety of real-world tasks. In this paper, we formulate a kernel level composition method to embed multiple local classifiers (kernels) into one kernel function, so as to obtain a more flexible data-dependent kernel. Since such composite kernels are composed by multiple local classifiers interpolated with several localizing gating functions, a specific learning process is also introduced in this paper to pre-determine their parameters. Experimental results are provided to validate two major perspectives of this paper. Firstly, the introduced learning process is effective to detect local information, which is essential for the parameter pre-determination of the localizing gating functions. Secondly, the proposed composite kernel has a capacity to improve classification performance.

UR - http://www.scopus.com/inward/record.url?scp=85007248132&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85007248132&partnerID=8YFLogxK

U2 - 10.1109/IJCNN.2016.7727696

DO - 10.1109/IJCNN.2016.7727696

M3 - Conference contribution

VL - 2016-October

SP - 3845

EP - 3850

BT - 2016 International Joint Conference on Neural Networks, IJCNN 2016

PB - Institute of Electrical and Electronics Engineers Inc.

ER -