Conditional Akaike information under covariate shift with application to small area estimation

Yuki Kawakubo; Shonosuke Sugasawa; Tatsuya Kubokawa

doi:10.1002/cjs.11354

Conditional Akaike information under covariate shift with application to small area estimation

Yuki Kawakubo, Shonosuke Sugasawa, Tatsuya Kubokawa

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

Abstract

In this study, we consider the problem of selecting explanatory variables of fixed effects in linear mixed models under covariate shift, which is when the values of covariates in the model for prediction differ from those in the model for observed data. We construct a variable selection criterion based on the conditional Akaike information introduced by Vaida & Blanchard (2005). We focus especially on covariate shift in small area estimation and demonstrate the usefulness of the proposed criterion. In addition, numerical performance is investigated through simulations, one of which is a design-based simulation using a real dataset of land prices.

Original language	English
Pages (from-to)	316-335
Number of pages	20
Journal	Canadian Journal of Statistics
Volume	46
Issue number	2
DOIs	https://doi.org/10.1002/cjs.11354
Publication status	Published - 2018 Jun
Externally published	Yes

Keywords

Akaike information criterion
conditional AIC
covariate shift
linear mixed model
MSC 2010: Primary 62J05
secondary 62P25
small area estimation

ASJC Scopus subject areas

Statistics and Probability
Statistics, Probability and Uncertainty

Access to Document

10.1002/cjs.11354

Cite this

@article{f7a8fd5c2e414fd68b69117b947da766,

title = "Conditional Akaike information under covariate shift with application to small area estimation",

abstract = "In this study, we consider the problem of selecting explanatory variables of fixed effects in linear mixed models under covariate shift, which is when the values of covariates in the model for prediction differ from those in the model for observed data. We construct a variable selection criterion based on the conditional Akaike information introduced by Vaida & Blanchard (2005). We focus especially on covariate shift in small area estimation and demonstrate the usefulness of the proposed criterion. In addition, numerical performance is investigated through simulations, one of which is a design-based simulation using a real dataset of land prices.",

keywords = "Akaike information criterion, conditional AIC, covariate shift, linear mixed model, MSC 2010: Primary 62J05, secondary 62P25, small area estimation",

author = "Yuki Kawakubo and Shonosuke Sugasawa and Tatsuya Kubokawa",

note = "Funding Information: The authors are grateful to the Editors and the anonymous referees for their valuable comments and helpful suggestions. The first and second authors are also grateful to Professor J. N. K. Rao for his helpful comment and for holding a seminar at Statistics Canada during their stay in Ottawa in 2015. This research was supported by Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science. The computational results were obtained using Ox version 6.21. Ox codes are available as the Supplementary Materials. Publisher Copyright: {\textcopyright} 2018 Statistical Society of Canada",

year = "2018",

month = jun,

doi = "10.1002/cjs.11354",

language = "English",

volume = "46",

pages = "316--335",

journal = "Canadian Journal of Statistics",

issn = "0319-5724",

publisher = "Statistical Society of Canada",

number = "2",

}

TY - JOUR

T1 - Conditional Akaike information under covariate shift with application to small area estimation

AU - Kawakubo, Yuki

AU - Sugasawa, Shonosuke

AU - Kubokawa, Tatsuya

N1 - Funding Information: The authors are grateful to the Editors and the anonymous referees for their valuable comments and helpful suggestions. The first and second authors are also grateful to Professor J. N. K. Rao for his helpful comment and for holding a seminar at Statistics Canada during their stay in Ottawa in 2015. This research was supported by Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science. The computational results were obtained using Ox version 6.21. Ox codes are available as the Supplementary Materials. Publisher Copyright: © 2018 Statistical Society of Canada

PY - 2018/6

Y1 - 2018/6

N2 - In this study, we consider the problem of selecting explanatory variables of fixed effects in linear mixed models under covariate shift, which is when the values of covariates in the model for prediction differ from those in the model for observed data. We construct a variable selection criterion based on the conditional Akaike information introduced by Vaida & Blanchard (2005). We focus especially on covariate shift in small area estimation and demonstrate the usefulness of the proposed criterion. In addition, numerical performance is investigated through simulations, one of which is a design-based simulation using a real dataset of land prices.

AB - In this study, we consider the problem of selecting explanatory variables of fixed effects in linear mixed models under covariate shift, which is when the values of covariates in the model for prediction differ from those in the model for observed data. We construct a variable selection criterion based on the conditional Akaike information introduced by Vaida & Blanchard (2005). We focus especially on covariate shift in small area estimation and demonstrate the usefulness of the proposed criterion. In addition, numerical performance is investigated through simulations, one of which is a design-based simulation using a real dataset of land prices.

KW - Akaike information criterion

KW - conditional AIC

KW - covariate shift

KW - linear mixed model

KW - MSC 2010: Primary 62J05

KW - secondary 62P25

KW - small area estimation

UR - http://www.scopus.com/inward/record.url?scp=85044239245&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85044239245&partnerID=8YFLogxK

U2 - 10.1002/cjs.11354

DO - 10.1002/cjs.11354

M3 - Article

AN - SCOPUS:85044239245

SN - 0319-5724

VL - 46

SP - 316

EP - 335

JO - Canadian Journal of Statistics

JF - Canadian Journal of Statistics

IS - 2

ER -

Conditional Akaike information under covariate shift with application to small area estimation

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this