Mixtures of regressions with predictor-dependent mixing proportions

D. S. Young; D. R. Hunter

doi:10.1016/j.csda.2010.04.002

Mixtures of regressions with predictor-dependent mixing proportions

D. S. Young, D. R. Hunter

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

48 Scopus citations

Abstract

We extend the standard mixture of linear regressions model by allowing the mixing proportions to be modeled nonparametrically as a function of the predictors. This framework allows for more flexibility in the modeling of the mixing proportions than the fully parametric mixture of experts model, which we also discuss. We present an EM-like algorithm for estimation of the new model. We also provide simulations demonstrating that our nonparametric approach can provide a better fit than the parametric approach in some instances and can serve to validate and thus reinforce the parametric approach in others. We also analyze and interpret two real data sets using the new method.

Original language	English
Pages (from-to)	2253-2266
Number of pages	14
Journal	Computational Statistics and Data Analysis
Volume	54
Issue number	10
DOIs	https://doi.org/10.1016/j.csda.2010.04.002
State	Published - Oct 1 2010

Bibliographical note

Funding Information:
We are grateful to two anonymous referees for numerous helpful comments during the preparation of this article. This research was supported by the National Science Foundation under Grant No. SES-0518772 .

Funding

We are grateful to two anonymous referees for numerous helpful comments during the preparation of this article. This research was supported by the National Science Foundation under Grant No. SES-0518772 .

Funders	Funder number
National Science Foundation (NSF)

Keywords

EM algorithms
Hierarchical mixture of experts
Mixture models
Mixtures of regressions

ASJC Scopus subject areas

Statistics and Probability
Computational Mathematics
Computational Theory and Mathematics
Applied Mathematics

Access to Document

10.1016/j.csda.2010.04.002

Cite this

@article{e4b2e55fd8a44046a99b88a5d177272d,

title = "Mixtures of regressions with predictor-dependent mixing proportions",

abstract = "We extend the standard mixture of linear regressions model by allowing the mixing proportions to be modeled nonparametrically as a function of the predictors. This framework allows for more flexibility in the modeling of the mixing proportions than the fully parametric mixture of experts model, which we also discuss. We present an EM-like algorithm for estimation of the new model. We also provide simulations demonstrating that our nonparametric approach can provide a better fit than the parametric approach in some instances and can serve to validate and thus reinforce the parametric approach in others. We also analyze and interpret two real data sets using the new method.",

keywords = "EM algorithms, Hierarchical mixture of experts, Mixture models, Mixtures of regressions",

author = "Young, {D. S.} and Hunter, {D. R.}",

note = "Funding Information: We are grateful to two anonymous referees for numerous helpful comments during the preparation of this article. This research was supported by the National Science Foundation under Grant No. SES-0518772 . ",

year = "2010",

month = oct,

day = "1",

doi = "10.1016/j.csda.2010.04.002",

language = "English",

volume = "54",

pages = "2253--2266",

number = "10",

}

TY - JOUR

T1 - Mixtures of regressions with predictor-dependent mixing proportions

AU - Young, D. S.

AU - Hunter, D. R.

N1 - Funding Information: We are grateful to two anonymous referees for numerous helpful comments during the preparation of this article. This research was supported by the National Science Foundation under Grant No. SES-0518772 .

PY - 2010/10/1

Y1 - 2010/10/1

N2 - We extend the standard mixture of linear regressions model by allowing the mixing proportions to be modeled nonparametrically as a function of the predictors. This framework allows for more flexibility in the modeling of the mixing proportions than the fully parametric mixture of experts model, which we also discuss. We present an EM-like algorithm for estimation of the new model. We also provide simulations demonstrating that our nonparametric approach can provide a better fit than the parametric approach in some instances and can serve to validate and thus reinforce the parametric approach in others. We also analyze and interpret two real data sets using the new method.

AB - We extend the standard mixture of linear regressions model by allowing the mixing proportions to be modeled nonparametrically as a function of the predictors. This framework allows for more flexibility in the modeling of the mixing proportions than the fully parametric mixture of experts model, which we also discuss. We present an EM-like algorithm for estimation of the new model. We also provide simulations demonstrating that our nonparametric approach can provide a better fit than the parametric approach in some instances and can serve to validate and thus reinforce the parametric approach in others. We also analyze and interpret two real data sets using the new method.

KW - EM algorithms

KW - Hierarchical mixture of experts

KW - Mixture models

KW - Mixtures of regressions

UR - http://www.scopus.com/inward/record.url?scp=77955275036&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77955275036&partnerID=8YFLogxK

U2 - 10.1016/j.csda.2010.04.002

DO - 10.1016/j.csda.2010.04.002

M3 - Article

AN - SCOPUS:77955275036

SN - 0167-9473

VL - 54

SP - 2253

EP - 2266

JO - Computational Statistics and Data Analysis

JF - Computational Statistics and Data Analysis

IS - 10

ER -

Mixtures of regressions with predictor-dependent mixing proportions

Abstract

Bibliographical note

Funding

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this