한국통계학회:학술대회논문집 (Proceedings of the Korean Statistical Society Conference)
- 한국통계학회 2005년도 춘계 학술발표회 논문집
- /
- Pages.215-219
- /
- 2005
Penalized Likelihood Regression: Fast Computation and Direct Cross-Validation
- Kim, Young-Ju (Department of Statistics, Kangwon National University) ;
- Gu, Chong (Department of Statistics, Purdue University)
- 발행 : 2005.05.20
초록
We consider penalized likelihood regression with exponential family responses. Parallel to recent development in Gaussian regression, the fast computation through asymptotically efficient low-dimensional approximations is explored, yielding algorithm that scales much better than the O(