Bayesian Inference on the Bimodality of the Generalized von Mises Distribution

Abstract This article introduces Bayesian inference on the bimodality of generalized von Mises (GvM) distribution for planar directions (Gatto and Jammalamadaka in Stat Methodol 4(3):341–353, 2007). The GvM is a flexible model that can be axial symmetric or asymmetric, unimodal bimodal. Two inferential approaches are analysed. first test null hypothesis Bayes factors obtained. second approach provides two-dimensional highest posterior density (HPD) credible set two parameters relevant to bimodality. Based identification parametric region associated with bimodality, inclusion HPD allows us infer underlying distribution. A particular implementation Metropolis–Hastings algorithm computation sets. Monte Carlo study reveals that, whenever samples generated under bimodal GvM, sets do clearly confirm