Title: Establishing a Bayesian nonparametric density estimation for biased circular data
Authors: Najmeh Nakhaeirad - University of Pretoria (South Africa) [presenting]
Andriette Bekker - University of Pretoria (South Africa)
Mohammad Arashi - Ferdowsi University of Mashhad (Iran)
Abstract: Circular data can be recorded with some errors in variables. This results in biased data. Routine directional methods fail to correctly model such data, therefore there is a demand to develop new estimation approaches. To pave the way for modelling such biased circular data, we first introduce a class of weighted distributions on the circle. We estimate the unknown forms of the distributions in the class by the kernel density estimation method. For posterior predictive density estimation, a Bayesian approach will be outlined and implemented. Numerical assessments, using the MCMC approach, support the findings, via simulation and real data analysis from a Bayesian nonparametric viewpoint.