Identifiability of asymmetric circular and cylindrical distributions

08/24/2019

∙

by   Yoichi Miyata, et al.

∙

0

∙

share

A new method to prove the identifiability of asymmetric circular and cylindrical distributions, which utilizes Teicher's approach, is proposed. We use the simultaneous Diophantine approximations and the trigonometric moments of circular random variables to check some conditions of the proposed method. We prove the identifiability of a general sine-skewed circular distribution including the sine-skewed von Mises and sine-skewed wrapped Cauchy distributions, and a cylindrical distribution combining the sine-skewed von Mises distribution on the circle and the Weibull distribution on the non-negative linear under suitable parameter spaces.