Orthogonality based modal empirical likelihood inferences for partially nonlinear models

This paper explored the effective empirical likelihood inferences for partially nonlinear models. By combining the modal regression method with orthogonal projection technology, a modal empirical likelihood-based estimation procedure was proposed. The proposed empirical likelihood approach retained...

Full description

Saved in:
Bibliographic Details
Published inAIMS mathematics Vol. 9; no. 7; pp. 18117 - 18133
Main Authors Lu, Jieqiong, Zhao, Peixin, Zhou, Xiaoshuang
Format Journal Article
LanguageEnglish
Published AIMS Press 01.01.2024
Subjects
empirical likelihood
modal regression
orthogonality estimation
partially nonlinear model
Online AccessGet full text

Cover

More Information
Summary:This paper explored the effective empirical likelihood inferences for partially nonlinear models. By combining the modal regression method with orthogonal projection technology, a modal empirical likelihood-based estimation procedure was proposed. The proposed empirical likelihood approach retained Wilk's theorem under mild conditions, and the confidence regions of model coefficients were constructed. Nonparametric and parametric components of the estimators were independent. Simulation results demonstrated that it is more robust and effective than the existing methods.
ISSN:2473-6988
2473-6988
DOI:10.3934/math.2024884