CMStatistics 2015: Start Registration
View Submission - CMStatistics
B0592
Title: Length dependent models for cylindrical data through sine skewed wrapped Cauchy distribution Authors:  Toshihiro Abe - Hosei University (Japan) [presenting]
Abstract: There has been an increased interest towards to the cylindrical distributions in recent years. The combination of direction and length, such as a pair of wind direction and its strength, can be regarded as the cylindrical data. Although some cylindrical models are well known, there are a few cylindrical models compared with the circular models. We propose another class of cylindrical distributions using circular distributions with certain dependence structure. The model has the simple normalizing constant. As a special case, we propose a cylindrical distribution whose conditional distribution is the sine-skewed wrapped Cauchy distribution. The advantages of such the selection are a general form of the trigonometric moments, simple generation of random number and explicit form of the mode.