Towards a unification of second-order theory for likelihood and marginal composite likelihood

• 成果类型: Article
• 署名作者: Lunardon, N.
• 署名单位: University of Padua
• 刊物名称: BIOMETRIKA
• ISSN/ISSBN: 0006-3444
• DOI: 10.1093/biomet/asv056
• 发表日期: 2016
• 页码: 225230
• 关键词: adjustment; inference; models

摘要

An adjustment for marginal composite likelihoods is derived to match the second-order theory of the likelihood when inference is for a vector-valued parameter in the absence of nuisance components. The adjustment overcomes the failure of Bartlett identities for marginal composite likelihoods and leads to a Bartlett-correctable marginal composite likelihood ratio statistic.