Failure of calibration is typical
From MaRDI portal
Foundations and philosophical topics in statistics (62A01) Bayesian problems; characterization of Bayes procedures (62C10) Bayesian inference (62F15) Classification and discrimination; cluster analysis (statistical aspects) (62H30) Topological data analysis (62R40) Applications of game theory (91A80)
Abstract: Schervish (1985b) showed that every forecasting system is noncalibrated for uncountably many data sequences that it might see. This result is strengthened here: from a topological point of view, failure of calibration is typical and calibration rare. Meanwhile, Bayesian forecasters are certain that they are calibrated---this invites worries about the connection between Bayesianism and rationality.
Recommendations
Cites work
- Algorithmic Learning Theory
- Asymptotic calibration
- Calibration-based empirical probability
- scientific article; zbMATH DE number 3128731 (Why is no real title available?)
- scientific article; zbMATH DE number 3976070 (Why is no real title available?)
- On calibration error of randomized forecasting algorithms
- On the Asymptotic Behavior of Bayes' Estimates in the Discrete Case
- Self-Calibrating Priors Do Not Exist
- The Well-Calibrated Bayesian
Cited in
(6)
This page was built for publication: Failure of calibration is typical
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2435755)