On choosing between two nonlinear models estimated robustly. some Monte Carlo evidence

Victor Aguirre-Torres, Andrew Ronald Gallant, Jorge Dorningueze

Research output: Contribution to journalArticle

Abstract

The article considers the problem of choosing between two (possibly) nonlinear models that have been fitted to the same data using M-estimation methods. An asymptotically normally distributed test statistic that takes into account the fact that the models are fitted robustly is given. The new procedure is compared with other test statistics using a Monte Carlo study. We found that the presence of a competitive model either in the null or the alternative hipothesis affects the distributional properties of the tests, and that in the case that the data contains outlying observations the new procedure had a significantly higher power than the rest of the tests.

Original languageEnglish (US)
Pages (from-to)179-200
Number of pages22
JournalCommunications in Statistics - Simulation and Computation
Volume18
Issue number1
DOIs
StatePublished - Jan 1 1989

Fingerprint

Test Statistic
Nonlinear Model
M-estimation
Monte Carlo Study
Statistics
High Power
Null
Alternatives
Model
Evidence
Observation

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Modeling and Simulation

Cite this

@article{3dc4d92bad754825bed061876e971aef,
title = "On choosing between two nonlinear models estimated robustly. some Monte Carlo evidence",
abstract = "The article considers the problem of choosing between two (possibly) nonlinear models that have been fitted to the same data using M-estimation methods. An asymptotically normally distributed test statistic that takes into account the fact that the models are fitted robustly is given. The new procedure is compared with other test statistics using a Monte Carlo study. We found that the presence of a competitive model either in the null or the alternative hipothesis affects the distributional properties of the tests, and that in the case that the data contains outlying observations the new procedure had a significantly higher power than the rest of the tests.",
author = "Victor Aguirre-Torres and Gallant, {Andrew Ronald} and Jorge Dorningueze",
year = "1989",
month = "1",
day = "1",
doi = "10.1080/03610918908812753",
language = "English (US)",
volume = "18",
pages = "179--200",
journal = "Communications in Statistics Part B: Simulation and Computation",
issn = "0361-0918",
publisher = "Taylor and Francis Ltd.",
number = "1",

}

On choosing between two nonlinear models estimated robustly. some Monte Carlo evidence. / Aguirre-Torres, Victor; Gallant, Andrew Ronald; Dorningueze, Jorge.

In: Communications in Statistics - Simulation and Computation, Vol. 18, No. 1, 01.01.1989, p. 179-200.

Research output: Contribution to journalArticle

TY - JOUR

T1 - On choosing between two nonlinear models estimated robustly. some Monte Carlo evidence

AU - Aguirre-Torres, Victor

AU - Gallant, Andrew Ronald

AU - Dorningueze, Jorge

PY - 1989/1/1

Y1 - 1989/1/1

N2 - The article considers the problem of choosing between two (possibly) nonlinear models that have been fitted to the same data using M-estimation methods. An asymptotically normally distributed test statistic that takes into account the fact that the models are fitted robustly is given. The new procedure is compared with other test statistics using a Monte Carlo study. We found that the presence of a competitive model either in the null or the alternative hipothesis affects the distributional properties of the tests, and that in the case that the data contains outlying observations the new procedure had a significantly higher power than the rest of the tests.

AB - The article considers the problem of choosing between two (possibly) nonlinear models that have been fitted to the same data using M-estimation methods. An asymptotically normally distributed test statistic that takes into account the fact that the models are fitted robustly is given. The new procedure is compared with other test statistics using a Monte Carlo study. We found that the presence of a competitive model either in the null or the alternative hipothesis affects the distributional properties of the tests, and that in the case that the data contains outlying observations the new procedure had a significantly higher power than the rest of the tests.

UR - http://www.scopus.com/inward/record.url?scp=84949329321&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84949329321&partnerID=8YFLogxK

U2 - 10.1080/03610918908812753

DO - 10.1080/03610918908812753

M3 - Article

VL - 18

SP - 179

EP - 200

JO - Communications in Statistics Part B: Simulation and Computation

JF - Communications in Statistics Part B: Simulation and Computation

SN - 0361-0918

IS - 1

ER -