Optimal Bayesian estimator under the reference prior in the von Mises distribution
Takenori Sakumura, Takemi Yanagimoto · Communications in Statistics - Simulation and Computation · 2025
A new Bayesian estimator is proposed for the von Mises distribution. It is found that the proposed estimator has desirable properties for estimating the concentration parameter of the von Mises distribution, which includes zero at the lower endpoint of the parameter space. This estimator is evaluated with several losses, and is found to be promising through comparison with existing estimators. A practical example demonstrates the possible usefulness of the proposed estimator.