Feasible generalized least squares using support vector regression

Steve Miller; Richard Startz

doi:10.1016/j.econlet.2018.12.001

Feasible generalized least squares using support vector regression

Steve Miller, Richard Startz

Applied Economics

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

8 Scopus citations

Abstract

We investigate semiparametric Feasible Generalized Least Squares using Support Vector Regression to estimate the conditional variance function. Monte Carlo results indicate the resulting estimator and an accompanying standard error correction offer substantially improved precision, nominal coverage rates, and shorter confidence intervals than Ordinary Least Squares with heteroskedasticity-consistent standard errors. Reductions in root mean squared error can be over 90% of those achievable when the form of heteroskedasticity is known.

Original language	English (US)
Pages (from-to)	28-31
Number of pages	4
Journal	Economics Letters
Volume	175
DOIs	https://doi.org/10.1016/j.econlet.2018.12.001
State	Published - Feb 2019

Bibliographical note

Funding Information:
We thank Marc Bellemare, Clément de Chaisemartin, Doug Steigerwald, and an anonymous referee for helpful comments and the Minnesota Supercomputing Institute for computational resources.

Keywords

Heteroskedasticity
Support vector regression
Weighted regression

Access

10.1016/j.econlet.2018.12.001

OpenUrl availability

Full text

Cite this

@article{c767e0440d024e2bb2e1585519106c65,

title = "Feasible generalized least squares using support vector regression",

abstract = "We investigate semiparametric Feasible Generalized Least Squares using Support Vector Regression to estimate the conditional variance function. Monte Carlo results indicate the resulting estimator and an accompanying standard error correction offer substantially improved precision, nominal coverage rates, and shorter confidence intervals than Ordinary Least Squares with heteroskedasticity-consistent standard errors. Reductions in root mean squared error can be over 90% of those achievable when the form of heteroskedasticity is known.",

keywords = "Heteroskedasticity, Support vector regression, Weighted regression",

author = "Steve Miller and Richard Startz",

note = "Funding Information: We thank Marc Bellemare, Cl{\'e}ment de Chaisemartin, Doug Steigerwald, and an anonymous referee for helpful comments and the Minnesota Supercomputing Institute for computational resources.",

year = "2019",

month = feb,

doi = "10.1016/j.econlet.2018.12.001",

language = "English (US)",

volume = "175",

pages = "28--31",

journal = "Economics Letters",

issn = "0165-1765",

publisher = "Elsevier",

}

TY - JOUR

T1 - Feasible generalized least squares using support vector regression

AU - Miller, Steve

AU - Startz, Richard

N1 - Funding Information: We thank Marc Bellemare, Clément de Chaisemartin, Doug Steigerwald, and an anonymous referee for helpful comments and the Minnesota Supercomputing Institute for computational resources.

PY - 2019/2

Y1 - 2019/2

N2 - We investigate semiparametric Feasible Generalized Least Squares using Support Vector Regression to estimate the conditional variance function. Monte Carlo results indicate the resulting estimator and an accompanying standard error correction offer substantially improved precision, nominal coverage rates, and shorter confidence intervals than Ordinary Least Squares with heteroskedasticity-consistent standard errors. Reductions in root mean squared error can be over 90% of those achievable when the form of heteroskedasticity is known.

AB - We investigate semiparametric Feasible Generalized Least Squares using Support Vector Regression to estimate the conditional variance function. Monte Carlo results indicate the resulting estimator and an accompanying standard error correction offer substantially improved precision, nominal coverage rates, and shorter confidence intervals than Ordinary Least Squares with heteroskedasticity-consistent standard errors. Reductions in root mean squared error can be over 90% of those achievable when the form of heteroskedasticity is known.

KW - Heteroskedasticity

KW - Support vector regression

KW - Weighted regression

UR - http://www.scopus.com/inward/record.url?scp=85058042806&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85058042806&partnerID=8YFLogxK

U2 - 10.1016/j.econlet.2018.12.001

DO - 10.1016/j.econlet.2018.12.001

M3 - Article

AN - SCOPUS:85058042806

SN - 0165-1765

VL - 175

SP - 28

EP - 31

JO - Economics Letters

JF - Economics Letters

ER -

Feasible generalized least squares using support vector regression

Abstract

Bibliographical note

Keywords

Access

OpenUrl availability

Other files and links

Fingerprint

Cite this