Game-theoretic probability combination with applications to resolving conflicts between statistical methods
From MaRDI portal
(Redirected from Publication:448957)
Abstract: In the typical analysis of a data set, a single method is selected for statistical reporting even when equally applicable methods yield very different results. Examples of equally applicable methods can correspond to those of different ancillary statistics in frequentist inference and of different prior distributions in Bayesian inference. More broadly, choices are made between parametric and nonparametric methods and between frequentist and Bayesian methods. Rather than choosing a single method, it can be safer, in a game-theoretic sense, to combine those that are equally appropriate in light of the available information. Since methods of combining subjectively assessed probability distributions are not objective enough for that purpose, this paper introduces a method of distribution combination that does not require any assignment of distribution weights. It does so by formalizing a hedging strategy in terms of a game between three players: nature, a statistician combining distributions, and a statistician refusing to combine distributions. The optimal move of the first statistician reduces to the solution of a simpler problem of selecting an estimating distribution that minimizes the Kullback-Leibler loss maximized over the plausible distributions to be combined. The resulting combined distribution is a linear combination of the most extreme of the distributions to be combined that are scientifically plausible. The optimal weights are close enough to each other that no extreme distribution dominates the others. The new methodology is illustrated by combining conflicting empirical Bayes methodologies in the context of gene expression data analysis.
Recommendations
Cites work
- scientific article; zbMATH DE number 3659450 (Why is no real title available?)
- scientific article; zbMATH DE number 3769296 (Why is no real title available?)
- scientific article; zbMATH DE number 3583699 (Why is no real title available?)
- scientific article; zbMATH DE number 1198604 (Why is no real title available?)
- scientific article; zbMATH DE number 720689 (Why is no real title available?)
- scientific article; zbMATH DE number 846108 (Why is no real title available?)
- scientific article; zbMATH DE number 2229942 (Why is no real title available?)
- scientific article; zbMATH DE number 3195782 (Why is no real title available?)
- scientific article; zbMATH DE number 3084669 (Why is no real title available?)
- A Bayesian perspective on the Bonferroni adjustment
- A Direct Approach to False Discovery Rates
- A general minimax result for relative entropy
- A new geometric capacity characterization of a discrete memoryless channel
- A predictive approach to measuring the strength of statistical evidence for single and multiple comparisons
- A source matching approach to finding minimax codes
- Calibration of \(\rho \) values for testing precise null hypotheses
- Coherent frequentism: a decision theory based on confidence sets
- Combining marginal probability distributions via minimization of weighted sum of Kullback-Leibler divergences
- Combining probability distributions: A critique and an annotated bibliography
- Comments on "A source matching approach to finding minimax codes" by Davisson, L. D. and Leon-Garcia, A.
- Confidence and Likelihood*
- Controlling the degree of caution in statistical inference with the Bayesian and frequentist approaches as opposite extremes
- Decision making with imprecise probabilistic information
- Elements of Information Theory
- Empirical Bayes Analysis of a Microarray Experiment
- Empirical Bayes interval estimates that are conditionally equal to unadjusted confidence intervals or to default prior credibility intervals
- Estimating the null distribution to adjust observed confidence levels for genome-scale screening
- Game theory, maximum entropy, minimum discrepancy and robust Bayesian decision theory
- Inference rules and inferential distributions
- Information and complexity in statistical modeling.
- Information theory. Coding theorems for discrete memoryless systems
- Large-scale inference. Empirical Bayes methods for estimation, testing, and prediction
- Marginalization and Linear Opinion Pools
- Principles of Statistical Inference
- Significance Tests in Parallel and in Series
- Size, power and false discovery rates
- Statistical decision theory and Bayesian analysis. 2nd ed
- The problem of regions
Cited in
(11)- A prior-free framework of coherent inference and its derivation of simple shrinkage estimators
- Correcting false discovery rates for their bias toward false positives
- Reporting Bayes factors or probabilities to decision makers of unknown loss functions
- Confidence distributions and empirical Bayes posterior distributions unified as distributions of evidential support
- Controlling the degree of caution in statistical inference with the Bayesian and frequentist approaches as opposite extremes
- Model fusion and multiple testing in the likelihood paradigm: shrinkage and evidence supporting a point null hypothesis
- Blending Bayesian and frequentist methods according to the precision of prior information with applications to hypothesis testing
- A further step for efficient corrections of inconsistent probabilistic data sets
- Inference after checking multiple Bayesian models for data conflict and applications to mitigating the influence of rejected priors
- Confidence distributions: a review
- Game-Theoretic Method of Estimating the Parameter of a Binomial Law from Results of One Observation
This page was built for publication: Game-theoretic probability combination with applications to resolving conflicts between statistical methods
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q448957)
