A composite-likelihood EM algorithm with importance sampling yields computationally feasible, asymptotically valid inference for the Poisson log-normal model on moderately large multivariate count datasets.
Note on information bias and efficiency of composite likelihood
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
Does the asymptotic variance of the maximum composite likelihood estimator of a parameter of interest always decrease when the nuisance parameters are known? Will a composite likelihood necessarily become more efficient by incorporating addi- tional independent component likelihoods, or by using component likelihoods with higher dimension? In this note we show through illustrative examples that the an- swer to both questions is no, and indeed the opposite direction might be observed. The role of information bias is highlighted to understand the occurrence of these paradoxical phenomenon.
fields
stat.CO 1years
2024 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Composite likelihood inference for the Poisson log-normal model
A composite-likelihood EM algorithm with importance sampling yields computationally feasible, asymptotically valid inference for the Poisson log-normal model on moderately large multivariate count datasets.