Bayesian Regression Analysis of Circular Data Using the Wrapping Approach
Ilaria Jansen · Utrecht University Repository (Utrecht University) · 2017
In this paper, a novel approach is introduced for circular regression based on the wrapping approach. Circular regression models are developed based on both the wrapped normal (WN) and wrapped Cauchy (WC) distribution. The scope of circular regression based on the wrapping approach is increased by making our methods applicable to models with an arbitrary number of continuous and dichotomous predictors. The resulting methods are validated by means of a simulation study, paying special attention to the differences between WN and WC. To demonstrate the use of our techniques, a regression model is fit to a real dataset from the field of Political Science. Diagnostic tools are provided to test reliability of the results and support inferences based on our methods