|
|
Royal Statistical Society |
| 1999 | |
|---|---|
|
5.00 p.m., 13 October |
PETER McCULLAGH (University of Chicago)
Invariance and factorial models The definition of a factorial model is extended to designs in which two or more factors have the same set of levels. Factorial models are first characterized by invariance under selection of levels. When applied to homologous factors, this same property identifies the statistically interesting subspaces, many of which are specified by symmetry, rotation and additivity. Appears (with discussion) in Journal of the Royal Statistical Society, Series B, 62(2), 209-256 (2000). |
| 2000 | |
|
5.00 p.m., 15 March |
MATTHEW STEPHENS AND PETER DONNELLY (University of Oxford)
Inference in molecular population genetics Modern molecular genetics generates extensive data which document the genetic variation in natural populations. Such data give rise to challenging statistical inference problems both for the underlying evolutionary parameters and for the demographic history of the population. This paper reviews recently introduced MCMC and importance sampling methods for approximating the relevant likelihood functions, and a novel importance sampling algorithm is developed. Comparison of these approaches suggests some general insights for computationally intensive inference in problems with high-dimensional missing data. Appears (with discussion) in Journal of the Royal Statistical Society, Series B, 62(4), 605-655 (2000). |