Cross-validation for selecting a model selection procedure

Yongli Zhang; Yuhong Yang

doi:10.1016/j.jeconom.2015.02.006

Cross-validation for selecting a model selection procedure

Yongli Zhang, Yuhong Yang

Statistics (Twin Cities)

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

196 Scopus citations

Abstract

While there are various model selection methods, an unanswered but important question is how to select one of them for data at hand. The difficulty is due to that the targeted behaviors of the model selection procedures depend heavily on uncheckable or difficult-to-check assumptions on the data generating process. Fortunately, cross-validation (CV) provides a general tool to solve this problem. In this work, results are provided on how to apply CV to consistently choose the best method, yielding new insights and guidance for potentially vast amount of application. In addition, we address several seemingly widely spread misconceptions on CV.

Original language	English (US)
Pages (from-to)	95-112
Number of pages	18
Journal	Journal of Econometrics
Volume	187
Issue number	1
DOIs	https://doi.org/10.1016/j.jeconom.2015.02.006
State	Published - Jul 1 2015

Bibliographical note

Funding Information:
We thank two anonymous referees, the Associate Editor and the Editor, Dr. Yacine Ait-Sahalia, for providing us with very insightful comments and valuable suggestions to improve the paper. The research of Yuhong Yang was partially supported by the United States National Science Foundation Grant DMS-1106576 .

Publisher Copyright:
© 2015 Elsevier B.V.

Keywords

Adaptive procedure selection
Cross-validation
Cross-validation paradox
Data splitting ratio
Information criterion
LASSO
MCP
SCAD

Publisher link

10.1016/j.jeconom.2015.02.006

Find It

Find It at U of M Libraries

Cite this

@article{e5391878d2f743b4b7abbb556f962309,

title = "Cross-validation for selecting a model selection procedure",

abstract = "While there are various model selection methods, an unanswered but important question is how to select one of them for data at hand. The difficulty is due to that the targeted behaviors of the model selection procedures depend heavily on uncheckable or difficult-to-check assumptions on the data generating process. Fortunately, cross-validation (CV) provides a general tool to solve this problem. In this work, results are provided on how to apply CV to consistently choose the best method, yielding new insights and guidance for potentially vast amount of application. In addition, we address several seemingly widely spread misconceptions on CV.",

keywords = "Adaptive procedure selection, Cross-validation, Cross-validation paradox, Data splitting ratio, Information criterion, LASSO, MCP, SCAD",

author = "Yongli Zhang and Yuhong Yang",

note = "Funding Information: We thank two anonymous referees, the Associate Editor and the Editor, Dr. Yacine Ait-Sahalia, for providing us with very insightful comments and valuable suggestions to improve the paper. The research of Yuhong Yang was partially supported by the United States National Science Foundation Grant DMS-1106576 . Publisher Copyright: {\textcopyright} 2015 Elsevier B.V.",

year = "2015",

month = jul,

day = "1",

doi = "10.1016/j.jeconom.2015.02.006",

language = "English (US)",

volume = "187",

pages = "95--112",

journal = "Journal of Econometrics",

issn = "0304-4076",

publisher = "Elsevier BV",

number = "1",

}

TY - JOUR

T1 - Cross-validation for selecting a model selection procedure

AU - Zhang, Yongli

AU - Yang, Yuhong

N1 - Funding Information: We thank two anonymous referees, the Associate Editor and the Editor, Dr. Yacine Ait-Sahalia, for providing us with very insightful comments and valuable suggestions to improve the paper. The research of Yuhong Yang was partially supported by the United States National Science Foundation Grant DMS-1106576 . Publisher Copyright: © 2015 Elsevier B.V.

PY - 2015/7/1

Y1 - 2015/7/1

N2 - While there are various model selection methods, an unanswered but important question is how to select one of them for data at hand. The difficulty is due to that the targeted behaviors of the model selection procedures depend heavily on uncheckable or difficult-to-check assumptions on the data generating process. Fortunately, cross-validation (CV) provides a general tool to solve this problem. In this work, results are provided on how to apply CV to consistently choose the best method, yielding new insights and guidance for potentially vast amount of application. In addition, we address several seemingly widely spread misconceptions on CV.

AB - While there are various model selection methods, an unanswered but important question is how to select one of them for data at hand. The difficulty is due to that the targeted behaviors of the model selection procedures depend heavily on uncheckable or difficult-to-check assumptions on the data generating process. Fortunately, cross-validation (CV) provides a general tool to solve this problem. In this work, results are provided on how to apply CV to consistently choose the best method, yielding new insights and guidance for potentially vast amount of application. In addition, we address several seemingly widely spread misconceptions on CV.

KW - Adaptive procedure selection

KW - Cross-validation

KW - Cross-validation paradox

KW - Data splitting ratio

KW - Information criterion

KW - LASSO

KW - MCP

KW - SCAD

UR - http://www.scopus.com/inward/record.url?scp=84929627804&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84929627804&partnerID=8YFLogxK

U2 - 10.1016/j.jeconom.2015.02.006

DO - 10.1016/j.jeconom.2015.02.006

M3 - Article

AN - SCOPUS:84929627804

VL - 187

SP - 95

EP - 112

JO - Journal of Econometrics

JF - Journal of Econometrics

SN - 0304-4076

IS - 1

ER -

Cross-validation for selecting a model selection procedure

Abstract

Bibliographical note

Keywords

Publisher link

Find It

Other files and links

Fingerprint

Cite this