Back to Search
Start Over
Data augmentation for non-Gaussian regression models using variance-mean mixtures
- Publication Year :
- 2011
-
Abstract
- We use the theory of normal variance-mean mixtures to derive a data-augmentation scheme for a class of common regularization problems. This generalizes existing theory on normal variance mixtures for priors in regression and classification. It also allows variants of the expectation-maximization algorithm to be brought to bear on a wider range of models than previously appreciated. We demonstrate the method on several examples, including sparse quantile regression and binary logistic regression. We also show that quasi-Newton acceleration can substantially improve the speed of the algorithm without compromising its robustness.<br />Comment: Added a discussion of quasi-Newton acceleration
- Subjects :
- Statistics - Methodology
Statistics - Computation
Subjects
Details
- Database :
- arXiv
- Publication Type :
- Report
- Accession number :
- edsarx.1103.5407
- Document Type :
- Working Paper