VVZ API is not affiliated with ETH Zurich. Data might be outdated or incorrect. Please view the official ETHZ Vorlesungsverzeichnis for binding information.
Asymptotic Statistics
Last Updated: 2026-02-05 15:29:11
Abstract
In this course, we treat selected topics from van der Vaart (1998).
Content
When comparing statistical procedures (tests, estimators, confidence intervals), one ideally would like to work with exact distributions. However, these are generally intractable, and might rely too heavily on perhaps unrealistic assumptions. Asymptotic statistics is about approximations of the distributions. Generally, the approximations are for large sample sizes, but also other approaches are possible. The advantage of asymptotic approaches is that on sees more clearly the common features of procedures (for example, likelihood ratio tests and score tests being asymptotically equivalent), and that certain desired properties can be more easily accomplished approximately (for example, unbiased estimators often do not exist, but asymptoticly unbiased ones are far more common). Moreover, asymptotic procedures can provide practical solutions, because by their general answers not every single case needs its own type of computations. Of course, there is also a disadvantage: the asymptotic approximations should be ``good enough". In this course, we treat selected topics from van der Vaart (1998).
Resources
Literature
van der Vaart, A.W. (1998). Asymptotic Statistics. Cambridge University Press
General Information
- Language
- English
- Levels
- BSC , MSC
Examination
- Type
- session examination
- Mode
- oral 20 minutes
Course Components
| Type | Title | Time & Place | Hours |
|---|---|---|---|
| lecture | Asymptotic Statistics |
|
2 h weekly |