Regularity,partial regularity,partial information process,for a filtered statistical model |
| |
Authors: | Jean Jacod |
| |
Institution: | (1) Laboratoire de Probabilités, Université Pierre et Marie Curie, Tour 56 (3e étage), 4, Place Jussieu, F-75252 Paris Cedex 05, France |
| |
Abstract: | Summary We define partial regularity for a filtered statistical (semi-parametric) model indexed by ![theta](/content/w65760tk0k543852/xxlarge952.gif) ![isin](/content/w65760tk0k543852/xxlarge8712.gif)
d
, as differentiability in a suitable sense of the partial likelihoods associated with a basic processX. Partial regularity turns out to be equivalent to some sort of differentiability in of the characteristics ofX. We also prove that regularity of the model implies partial regularity, and we define a partial information process , which is smaller than the complete information process. We apply these results to obtain a generalization of Cramer-Rao inequality, and to prove that partial likelihood processes are optimal among all quasi-likelihood processes which are stochastic integrals with respect to the basic processX. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|