Abstract
Mathematical definitions are given of the observed and expected Fisher information matrix. A description is given of the role of Fisher information as a lower bound on the variance of unbiased parameter estimates (Cramer–Rao inequality) and as the asymptotic variance of the maximum likelihood estimate.
| Original language | English |
|---|---|
| Title of host publication | Encyclopedia of Biostatistics |
| Subtitle of host publication | Armitage Enc Biostats 2e |
| Publisher | wiley |
| Pages | 1 |
| Number of pages | 1 |
| ISBN (Electronic) | 9780470011812 |
| ISBN (Print) | 9780470849071 |
| DOIs | |
| State | Published - 1 Jan 2006 |
Bibliographical note
Publisher Copyright:© 2005 John Wiley & Sons, Ltd.
Keywords
- likelihood
- variance