EUVIMEDAtklāt EUVIMED
Open access · CC BY 4.0

Fast Estimation and Valid Statistical Inference for Mixed‐Effect Location‐Scale Models Using Variational Inference

Brian Ping-Huan Wu iD, Donald Hedeker

Statistics in Medicine · 2026-06 · 45 (13-14)

DOI: 10.1002/sim.70640 ↗

At a glance

Journal article in Statistics in Medicine, published online on 2026-06-17. Written by Brian Ping-Huan Wu and 1 co-authors. Institution: Department of Statistics, University of Chicago, Chicago, Illinois, USA. Major MEDLINE topics: Algorithms, Models, Statistical. 39 references. The full text is freely available (open access, CC BY 4.0). Funded by: NIH HHS, National Institutes of Health.

AbstraktsSubjectsBibliogrāfiskie datiCiteRelated studies

Abstract (original language)

With the development of modern technology, richer and more intensive longitudinal data are available for researchers to study more complex model structures. The Mixed‐Effect Location‐Scale Model (MELS) is useful for modeling the variance components in such data. However, fitting these models is computationally demanding using standard Maximum Likelihood estimation (MLE) or Markov Chain Monte Carlo (MCMC) methods. This paper introduces a fast, deterministic Variational Message Passing (VMP) algorithm for fitting MELS models. We derive the updates for the non‐conjugate components using a simplified Laplace approximation and employ a robust M‐estimation framework to construct sandwich estimators for valid frequentist inference. Simulation studies confirm our estimator is accurate and consistent. A real data example shows that our algorithm achieves point and interval estimates comparable to MLE in seconds, outperforming MLE (minutes) and MCMC (hours). This MELS‐VMP algorithm provides a computationally efficient and reliable alternative for model selection and inference on intensive longitudinal datasets.

Subjects

MeSH terms (MEDLINE)

AlgorithmsModels, StatisticalLikelihood FunctionsComputer SimulationMarkov ChainsMonte Carlo MethodHumansLongitudinal StudiesData Interpretation, Statistical

Author keywords

longitudinal datamixed effect modelsrobust standard errorsvariational inference

Authors and institutions

  • Department of Statistics, University of Chicago, Chicago, Illinois, USA.
  • Department of Public Health Sciences, University of Chicago, Chicago, Illinois, USA.

ORCID profiles: Brian Ping-Huan Wu

Funding

  • NIH HHS
  • National Institutes of Health

Cite

APA

Wu, B. P., & Hedeker, D. (2026). Fast Estimation and Valid Statistical Inference for Mixed‐Effect Location‐Scale Models Using Variational Inference. Statistics in Medicine, 45(13-14). https://doi.org/10.1002/sim.70640

Vancouver

Wu BP, Hedeker D. Fast Estimation and Valid Statistical Inference for Mixed‐Effect Location‐Scale Models Using Variational Inference. Statistics in Medicine. 2026;45(13-14). doi:10.1002/sim.70640

BibTeX
@article{Wu2026,
  author = {Wu, Brian Ping-Huan and Hedeker, Donald},
  title = {Fast Estimation and Valid Statistical Inference for Mixed‐Effect Location‐Scale Models Using Variational Inference},
  journal = {Statistics in Medicine},
  year = {2026},
  volume = {45},
  number = {13-14},
  doi = {10.1002/sim.70640}
}
Start for free →