A general framework for frequentist model averaging

Priyam Mitra, Heng Lian, Ritwik Mitra, Hua Liang, Minge Xie

Research output: Contribution to journalArticle

Abstract

Model selection strategies have been routinely employed to determine a model for data analysis in statistics, and further study and inference then often proceed as though the selected model were the true model that were known a priori. Model averaging approaches, on the other hand, try to combine estimators for a set of candidate models. Specifically, instead of deciding which model is the `right' one, a model averaging approach suggests to fit a set of candidate models and average over the estimators using data adaptive weights. In this paper we establish a general frequentist model averaging framework that does not set any restrictions on the set of candidate models. It broadens the scope of the existing methodologies under the frequentist model averaging development. Assuming the data is from an unknown model, we derive the model averaging estimator and study its limiting distributions and related predictions while taking possible modeling biases into account. We propose a set of optimal weights to combine the individual estimators so that the expected mean squared error of the average estimator is minimized. Simulation studies are conducted to compare the performance of the estimator with that of the existing methods. The results show the benefits of the proposed approach over traditional model selection approaches as well as existing model averaging methods.

Original languageEnglish (US)
Pages (from-to)205-226
Number of pages22
JournalScience China Mathematics
Volume62
Issue number2
DOIs
StatePublished - Feb 1 2019

Fingerprint

Model Averaging
Estimator
Model Selection
Model
Framework
Averaging Method
Limiting Distribution
Mean Squared Error
Data analysis
Simulation Study
Restriction
Statistics
Unknown
Methodology
Prediction

Cite this

Mitra, Priyam ; Lian, Heng ; Mitra, Ritwik ; Liang, Hua ; Xie, Minge. / A general framework for frequentist model averaging. In: Science China Mathematics. 2019 ; Vol. 62, No. 2. pp. 205-226.
@article{552da8cef0164bc986d559e88ac85590,
title = "A general framework for frequentist model averaging",
abstract = "Model selection strategies have been routinely employed to determine a model for data analysis in statistics, and further study and inference then often proceed as though the selected model were the true model that were known a priori. Model averaging approaches, on the other hand, try to combine estimators for a set of candidate models. Specifically, instead of deciding which model is the `right' one, a model averaging approach suggests to fit a set of candidate models and average over the estimators using data adaptive weights. In this paper we establish a general frequentist model averaging framework that does not set any restrictions on the set of candidate models. It broadens the scope of the existing methodologies under the frequentist model averaging development. Assuming the data is from an unknown model, we derive the model averaging estimator and study its limiting distributions and related predictions while taking possible modeling biases into account. We propose a set of optimal weights to combine the individual estimators so that the expected mean squared error of the average estimator is minimized. Simulation studies are conducted to compare the performance of the estimator with that of the existing methods. The results show the benefits of the proposed approach over traditional model selection approaches as well as existing model averaging methods.",
author = "Priyam Mitra and Heng Lian and Ritwik Mitra and Hua Liang and Minge Xie",
year = "2019",
month = "2",
day = "1",
doi = "https://doi.org/10.1007/s11425-018-9403-x",
language = "English (US)",
volume = "62",
pages = "205--226",
journal = "Science China Mathematics",
issn = "1674-7283",
publisher = "Science in China Press",
number = "2",

}

A general framework for frequentist model averaging. / Mitra, Priyam; Lian, Heng; Mitra, Ritwik; Liang, Hua; Xie, Minge.

In: Science China Mathematics, Vol. 62, No. 2, 01.02.2019, p. 205-226.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A general framework for frequentist model averaging

AU - Mitra, Priyam

AU - Lian, Heng

AU - Mitra, Ritwik

AU - Liang, Hua

AU - Xie, Minge

PY - 2019/2/1

Y1 - 2019/2/1

N2 - Model selection strategies have been routinely employed to determine a model for data analysis in statistics, and further study and inference then often proceed as though the selected model were the true model that were known a priori. Model averaging approaches, on the other hand, try to combine estimators for a set of candidate models. Specifically, instead of deciding which model is the `right' one, a model averaging approach suggests to fit a set of candidate models and average over the estimators using data adaptive weights. In this paper we establish a general frequentist model averaging framework that does not set any restrictions on the set of candidate models. It broadens the scope of the existing methodologies under the frequentist model averaging development. Assuming the data is from an unknown model, we derive the model averaging estimator and study its limiting distributions and related predictions while taking possible modeling biases into account. We propose a set of optimal weights to combine the individual estimators so that the expected mean squared error of the average estimator is minimized. Simulation studies are conducted to compare the performance of the estimator with that of the existing methods. The results show the benefits of the proposed approach over traditional model selection approaches as well as existing model averaging methods.

AB - Model selection strategies have been routinely employed to determine a model for data analysis in statistics, and further study and inference then often proceed as though the selected model were the true model that were known a priori. Model averaging approaches, on the other hand, try to combine estimators for a set of candidate models. Specifically, instead of deciding which model is the `right' one, a model averaging approach suggests to fit a set of candidate models and average over the estimators using data adaptive weights. In this paper we establish a general frequentist model averaging framework that does not set any restrictions on the set of candidate models. It broadens the scope of the existing methodologies under the frequentist model averaging development. Assuming the data is from an unknown model, we derive the model averaging estimator and study its limiting distributions and related predictions while taking possible modeling biases into account. We propose a set of optimal weights to combine the individual estimators so that the expected mean squared error of the average estimator is minimized. Simulation studies are conducted to compare the performance of the estimator with that of the existing methods. The results show the benefits of the proposed approach over traditional model selection approaches as well as existing model averaging methods.

UR - http://www.scopus.com/inward/record.url?scp=85060220825&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85060220825&partnerID=8YFLogxK

U2 - https://doi.org/10.1007/s11425-018-9403-x

DO - https://doi.org/10.1007/s11425-018-9403-x

M3 - Article

VL - 62

SP - 205

EP - 226

JO - Science China Mathematics

JF - Science China Mathematics

SN - 1674-7283

IS - 2

ER -