Skew T and Semiparametric Empirical Likelihoods Versus Parametric Robust Likelihood

• Activity Number: 193
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Wei-Cheng Hsiao* and Tsung-Shan Tsou
• Companies: Institute of Statistical Science, Academia Sinica and National Central University
• Keywords: skew t distribution ; empirical likelihood ; robust likelihood ; model misspecification

Abstract

Likelihood is certainly one of the most important entities for inference. The skew t is a popular distribution that researchers employ to analyze asymmetric data. Our findings, however, reveal that the skew t model might not be able to fulfill its intended goal. The empirical likelihood is a distribution-free likelihood approach. The adjusted empirical likelihood functions exhibit many properties of the parametric likelihood functions and the AEL* (Liu and Chen, 2010) has been shown to outperform its predecessor. The robust likelihood introduced by Royall and Tsou (2003) is a parametric and yet robust approach. Our study shows that the parametric robust likelihood approach might be a better choice than AEL*.