Corpus ID: 104292027

Meta-analysis of Bayesian analyses.

  title={Meta-analysis of Bayesian analyses.},
  author={P. Blomstedt and Diego Mesquita and Jarno Lintusaari and Tuomas Sivula and J. Corander and Samuel Kaski},
  journal={arXiv: Methodology},
  • P. Blomstedt, Diego Mesquita, +3 authors Samuel Kaski
  • Published 2019
  • Mathematics
  • arXiv: Methodology
  • Meta-analysis aims to combine results from multiple related statistical analyses. While the natural outcome of a Bayesian analysis is a posterior distribution, Bayesian meta-analyses traditionally combine analyses summarized as point estimates, often limiting distributional assumptions. In this paper, we develop a framework for combining posterior distributions, which builds on standard Bayesian inference, but using distributions instead of data points as observations. We show that the… CONTINUE READING
    2 Citations

    Figures from this paper

    A numerically stable algorithm for integrating Bayesian models using Markov melding.
    • PDF
    Statistical Model Aggregation via Parameter Matching
    • 7
    • PDF


    A re-evaluation of random-effects meta-analysis
    • 1,275
    • PDF
    Fully Bayesian hierarchical modelling in two stages, with application to meta-analysis
    • 47
    Confidence Distributions and a Unifying Framework for Meta-Analysis
    • 99
    • PDF
    Meta-analysis: formulating, evaluating, combining, and reporting.
    • S. Normand
    • Computer Science, Medicine
    • Statistics in medicine
    • 1999
    • 862
    • PDF
    Bayesian statistics without tears: A sampling-resampling perspective
    • 932
    • PDF
    Functional regression approximate Bayesian computation for Gaussian process density estimation
    • 8
    • PDF
    Post-Inference Prior Swapping
    • 7
    • PDF
    Confidence and likelihood
    • 151
    • PDF