CMStatistics 2017: Start Registration
View Submission - CMStatistics
B1530
Title: A study on a general family of asymmetric distributions Authors:  Irene Gijbels - Katholieke Universiteit Leuven (Belgium)
Md Rezaul Karim - Katholieke Universiteit Leuven (Belgium) [presenting]
Anneleen Verhasselt - Hasselt University (Belgium)
Abstract: A general family of asymmetric distributions is studied in which the location parameter is a specific quantile of the distribution (which is the main attraction of this family). We present the resulting special cases of asymmetric normal, asymmetric student-$t$ and asymmetric logistic distributions and investigate their properties. We discuss parameter estimation procedures by using method of moments and maximum likelihood estimation. We briefly discuss advantages and disadvantages of the estimation methods, and establish their asymptotic behaviour. The proposed family of asymmetric distributions can be applied in a conditional setting, allowing to study regression quantiles in a likelihood framework. We consider both parametric and semi-parametric estimation in quantile regression settings.