Proceedings of the Korean Statistical Society Conference (한국통계학회:학술대회논문집)
- 2005.05a
- /
- Pages.109-112
- /
- 2005
Asymmetric robust quasi-likelihood
- Lee, Yoon-Dong (Applied Statistics, Konkuk University) ;
- Choi, Hye-Mi (Applied Mathematics, KAIST)
- Published : 2005.05.20
Abstract
The robust quasi-likelihood (RQL) proposed by Cantoni & Ronchetti (2001) is a robust version of quasi-likelihood. They adopted Huber function to increase the resistance of the RQL estimator to the outliers. They considered the Huber function only of symmetric type. We extend the class of Huber function to include asymmetric types, and derived a method to find the optimal asymmetric one.
Keywords