en
Scientific article
English

Composite likelihood inference by nonparametric saddlepoint tests

Published inComputational statistics & data analysis, vol. 79, p. 80-90
Publication date2014
Abstract

The class of composite likelihood functions provides a flexible and powerful toolkit to carry out approximate inference for complex statistical models when the full likelihood is either impossible to specify or unfeasible to compute. However, the strength of the composite likelihood approach is dimmed when considering hypothesis testing about a multidimensional parameter because the finite sample behavior of likelihood ratio, Wald, and score-type test statistics is tied to the Godambe information matrix. Consequently, inaccurate estimates of the Godambe information translate in inaccurate p-values. The approach based on a fully nonparametric saddlepoint test statistic derived from the composite score functions is shown to achieve accurate inference. The proposed statistic is asymptotically chi-squared distributed up to a relative error of second order and does not depend on the Godambe information. The validity of the method is demonstrated through simulation studies.

Keywords
  • Empirical likelihood methods
  • Godambe information
  • Likelihood ratio adjustments
  • Nonparametric inference
  • Pairwise likelihood
  • Relative error Robust tests Saddlepoint test Small sample inference
Citation (ISO format)
LUNARDON, Nicola, RONCHETTI, Elvezio. Composite likelihood inference by nonparametric saddlepoint tests. In: Computational statistics & data analysis, 2014, vol. 79, p. 80–90. doi: 10.1016/j.csda.2014.05.014
Main files (1)
Article (Published version)
accessLevelRestricted
Identifiers
ISSN of the journal0167-9473
583views
5downloads

Technical informations

Creation01/27/2015 5:08:00 PM
First validation01/27/2015 5:08:00 PM
Update time03/14/2023 10:47:01 PM
Status update03/14/2023 10:47:01 PM
Last indexation05/02/2024 3:58:21 PM
All rights reserved by Archive ouverte UNIGE and the University of GenevaunigeBlack