B0882
Title: An extension of Johnson-Wehrly-type cylindrical distributions
Authors: Yoichi Miyata - Takasaki City University of Economics (Japan) [presenting]
Abstract: The Johnson-Wehrly-type cylindrical distributions in which the circular marginal distributions are the sine-skewed wrapped Cauchy and the conditional distributions of a linear random variable are the Weibull are known to be identifiable, and have simple normalizing constants and easy random number generation algorithms. Extending the models, we propose a new family of cylindrical distributions in which the circular marginal distributions are the extended sine-skewed wrapped Cauchy. The proposed distributions also have simple normalizing constants and easy random number generation algorithms, and allow marginal distributions of a circular random variable to be skewed more significantly than that of the above models. Numerical examples of fitting this model to the Periwinkle data and a hidden Markov model with each component being the proposed distribution are shown.