Convex support vector regression

Zhiqiang Liao; Sheng Dai; Timo Kuosmanen

doi:10.1016/j.ejor.2023.05.009

Convex support vector regression

Zhiqiang Liao, Sheng Dai^*, Timo Kuosmanen

^*Corresponding author for this work

University of Turku

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

28 Citations (Scopus)

23 Downloads (Pure)

Abstract

Nonparametric regression subject to convexity or concavity constraints is increasingly popular in economics, finance, operations research, machine learning, and statistics. However, the conventional convex regression based on the least squares loss function often suffers from overfitting and outliers. This paper proposes to address these two issues by introducing the convex support vector regression (CSVR) method, which effectively combines the key elements of convex regression and support vector regression. Numerical experiments demonstrate the performance of CSVR in prediction accuracy and robustness that compares favorably with other state-of-the-art methods.

Original language	English
Pages (from-to)	858-870
Number of pages	13
Journal	European Journal of Operational Research
Volume	313
Issue number	3
Early online date	9 May 2023
DOIs	https://doi.org/10.1016/j.ejor.2023.05.009
Publication status	Published - Mar 2024
MoE publication type	A1 Journal article-refereed

Funding

The authors would like to thank the three anonymous reviewers for their helpful comments. We acknowledge the computational resources provided by the Aalto Science-IT project. Zhiqiang Liao gratefully acknowledges financial support from the Foundation for Economic Education (Liikesivistysrahasto) [grant no. 210038] and the Jenny and Antti Wihuri Foundation [grant no. 00220201]. Sheng Dai gratefully acknowledges financial support from the Foundation for Economic Education (Liikesivistysrahasto) [grant no. 220074] and the OP Group Research Foundation [grant no. 20230008].

Keywords

Convex regression
Overfitting
Regularization
Robustness and sensitivity analysis
Support vector regression

Access to Document

10.1016/j.ejor.2023.05.009Licence: CC BY

Liao-et-al_Convex-support-vector-regression_2023Final published version, 4.55 MBLicence: CC BY

Cite this

@article{34d277825e75421a8720e2665989144b,

title = "Convex support vector regression",

abstract = "Nonparametric regression subject to convexity or concavity constraints is increasingly popular in economics, finance, operations research, machine learning, and statistics. However, the conventional convex regression based on the least squares loss function often suffers from overfitting and outliers. This paper proposes to address these two issues by introducing the convex support vector regression (CSVR) method, which effectively combines the key elements of convex regression and support vector regression. Numerical experiments demonstrate the performance of CSVR in prediction accuracy and robustness that compares favorably with other state-of-the-art methods.",

keywords = "Convex regression, Overfitting, Regularization, Robustness and sensitivity analysis, Support vector regression",

author = "Zhiqiang Liao and Sheng Dai and Timo Kuosmanen",

note = "Funding Information: The authors would like to thank the three anonymous reviewers for their helpful comments. We acknowledge the computational resources provided by the Aalto Science-IT project. Zhiqiang Liao gratefully acknowledges financial support from the Foundation for Economic Education (Liikesivistysrahasto) [grant no. 210038] and the Jenny and Antti Wihuri Foundation [grant no. 00220201]. Sheng Dai gratefully acknowledges financial support from the Foundation for Economic Education (Liikesivistysrahasto) [grant no. 220074] and the OP Group Research Foundation [grant no. 20230008]. Publisher Copyright: {\textcopyright} 2023 The Author(s)",

year = "2024",

month = mar,

doi = "10.1016/j.ejor.2023.05.009",

language = "English",

volume = "313",

pages = "858--870",

journal = "European Journal of Operational Research",

issn = "0377-2217",

publisher = "Elsevier",

number = "3",

}

TY - JOUR

T1 - Convex support vector regression

AU - Liao, Zhiqiang

AU - Dai, Sheng

AU - Kuosmanen, Timo

N1 - Funding Information: The authors would like to thank the three anonymous reviewers for their helpful comments. We acknowledge the computational resources provided by the Aalto Science-IT project. Zhiqiang Liao gratefully acknowledges financial support from the Foundation for Economic Education (Liikesivistysrahasto) [grant no. 210038] and the Jenny and Antti Wihuri Foundation [grant no. 00220201]. Sheng Dai gratefully acknowledges financial support from the Foundation for Economic Education (Liikesivistysrahasto) [grant no. 220074] and the OP Group Research Foundation [grant no. 20230008]. Publisher Copyright: © 2023 The Author(s)

PY - 2024/3

Y1 - 2024/3

N2 - Nonparametric regression subject to convexity or concavity constraints is increasingly popular in economics, finance, operations research, machine learning, and statistics. However, the conventional convex regression based on the least squares loss function often suffers from overfitting and outliers. This paper proposes to address these two issues by introducing the convex support vector regression (CSVR) method, which effectively combines the key elements of convex regression and support vector regression. Numerical experiments demonstrate the performance of CSVR in prediction accuracy and robustness that compares favorably with other state-of-the-art methods.

AB - Nonparametric regression subject to convexity or concavity constraints is increasingly popular in economics, finance, operations research, machine learning, and statistics. However, the conventional convex regression based on the least squares loss function often suffers from overfitting and outliers. This paper proposes to address these two issues by introducing the convex support vector regression (CSVR) method, which effectively combines the key elements of convex regression and support vector regression. Numerical experiments demonstrate the performance of CSVR in prediction accuracy and robustness that compares favorably with other state-of-the-art methods.

KW - Convex regression

KW - Overfitting

KW - Regularization

KW - Robustness and sensitivity analysis

KW - Support vector regression

UR - https://www.scopus.com/pages/publications/85160075894

U2 - 10.1016/j.ejor.2023.05.009

DO - 10.1016/j.ejor.2023.05.009

M3 - Article

AN - SCOPUS:85160075894

SN - 0377-2217

VL - 313

SP - 858

EP - 870

JO - European Journal of Operational Research

JF - European Journal of Operational Research

IS - 3

ER -

Convex support vector regression

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Science-IT

Cite this

Convex support vector regression

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Equipment

Science-IT

Cite this