A Bayesian analysis of directional data using the projected normal distribution

Abstract

This paper presents a Bayesian analysis of the projected normal distribution, which is a flexible and useful distribution for the analysis of directional data. We obtain samples from the posterior distribution using the Gibbs sampler after the introduction of suitably chosen latent variables. The procedure is illustrated using simulated data as well as a real data set previously analysed in the literature.

Keywords:

• Circular data
• Gibbs sampler
• latent variables
• radial projection
• spherical data

Acknowledgements

The work of the second author was supported by Grant 32256-E from CONACYT, Mexico, and by the Sistema Nacional de Investigadores, Mexico.