Principal weighted support vector machines for sufficient dimension reduction in binary classification

Seung Jun Shin; Yichao Wu; Hao Helen Zhang; Yufeng Liu

doi:10.1093/biomet/asw057

Principal weighted support vector machines for sufficient dimension reduction in binary classification

Seung Jun Shin, Yichao Wu, Hao Helen Zhang, Yufeng Liu

Department of Statistics

Research output: Contribution to journal › Article › peer-review

28 Citations (Scopus)

Abstract

Sufficient dimension reduction is popular for reducing data dimensionality without stringent model assumptions. However, most existing methods may work poorly for binary classification. For example, sliced inverse regression (Li, 1991) can estimate at most one direction if the response is binary. In this paper we propose principal weighted support vector machines, a unified framework for linear and nonlinear sufficient dimension reduction in binary classification. Its asymptotic properties are studied, and an efficient computing algorithm is proposed. Numerical examples demonstrate its performance in binary classification.

Original language	English
Pages (from-to)	67-81
Number of pages	15
Journal	Biometrika
Volume	104
Issue number	1
DOIs	https://doi.org/10.1093/biomet/asw057
Publication status	Published - 2017 Mar 1

Keywords

Fisher consistency
Hyperplane alignment
Reproducing kernel Hilbert space
Weighted support vector machine

ASJC Scopus subject areas

Statistics and Probability
Mathematics(all)
Agricultural and Biological Sciences (miscellaneous)
Agricultural and Biological Sciences(all)
Statistics, Probability and Uncertainty
Applied Mathematics

Access to Document

10.1093/biomet/asw057

Cite this

@article{96f7bf2a3ed44012a2172c4b6a008f5f,

title = "Principal weighted support vector machines for sufficient dimension reduction in binary classification",

abstract = "Sufficient dimension reduction is popular for reducing data dimensionality without stringent model assumptions. However, most existing methods may work poorly for binary classification. For example, sliced inverse regression (Li, 1991) can estimate at most one direction if the response is binary. In this paper we propose principal weighted support vector machines, a unified framework for linear and nonlinear sufficient dimension reduction in binary classification. Its asymptotic properties are studied, and an efficient computing algorithm is proposed. Numerical examples demonstrate its performance in binary classification.",

keywords = "Fisher consistency, Hyperplane alignment, Reproducing kernel Hilbert space, Weighted support vector machine",

author = "Shin, {Seung Jun} and Yichao Wu and Zhang, {Hao Helen} and Yufeng Liu",

note = "Funding Information: Our research is partially supported by the National Institutes of Health, the National Science Foundation and the National Research Foundation of Korea. Publisher Copyright: {\textcopyright} 2017 Biometrika Trust.",

year = "2017",

month = mar,

day = "1",

doi = "10.1093/biomet/asw057",

language = "English",

volume = "104",

pages = "67--81",

journal = "Biometrika",

issn = "0006-3444",

publisher = "Oxford University Press",

number = "1",

}

TY - JOUR

T1 - Principal weighted support vector machines for sufficient dimension reduction in binary classification

AU - Shin, Seung Jun

AU - Wu, Yichao

AU - Zhang, Hao Helen

AU - Liu, Yufeng

N1 - Funding Information: Our research is partially supported by the National Institutes of Health, the National Science Foundation and the National Research Foundation of Korea. Publisher Copyright: © 2017 Biometrika Trust.

PY - 2017/3/1

Y1 - 2017/3/1

N2 - Sufficient dimension reduction is popular for reducing data dimensionality without stringent model assumptions. However, most existing methods may work poorly for binary classification. For example, sliced inverse regression (Li, 1991) can estimate at most one direction if the response is binary. In this paper we propose principal weighted support vector machines, a unified framework for linear and nonlinear sufficient dimension reduction in binary classification. Its asymptotic properties are studied, and an efficient computing algorithm is proposed. Numerical examples demonstrate its performance in binary classification.

AB - Sufficient dimension reduction is popular for reducing data dimensionality without stringent model assumptions. However, most existing methods may work poorly for binary classification. For example, sliced inverse regression (Li, 1991) can estimate at most one direction if the response is binary. In this paper we propose principal weighted support vector machines, a unified framework for linear and nonlinear sufficient dimension reduction in binary classification. Its asymptotic properties are studied, and an efficient computing algorithm is proposed. Numerical examples demonstrate its performance in binary classification.

KW - Fisher consistency

KW - Hyperplane alignment

KW - Reproducing kernel Hilbert space

KW - Weighted support vector machine

UR - http://www.scopus.com/inward/record.url?scp=85019930767&partnerID=8YFLogxK

U2 - 10.1093/biomet/asw057

DO - 10.1093/biomet/asw057

M3 - Article

AN - SCOPUS:85019930767

SN - 0006-3444

VL - 104

SP - 67

EP - 81

JO - Biometrika

JF - Biometrika

IS - 1

ER -

Principal weighted support vector machines for sufficient dimension reduction in binary classification

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this