Credits
12 points (18 ECTS credits)

Lecturer
Björn Holmquist
phone; +46 46 222 8926
e-mail: Bjorn.Holmquist@stat.lu.se
office: 38
Tel no: (046) 222 8926

Registration:
The secretary at the department, Lena Somogyi, phone: +46 46 2228921, email: lena@stat.lu.se.

Meetings
The course meet once or twice a week during end of  September 2000 - February 2001.

Examination
Take home examination and oral test.

Course contents
Models for observations and principles for statistical inference, sufficient and ancillary statistics, equivariance and invariance.

Estimation: Bias versus variance, Bayesian estimators, inequalities for minimal variance, asymptotic theory for Maximum Likelihood estimators, partial likelihood, robust estimators, applications in non-parametric regression, finite populations, and case-control studies.

Hypothesis testing: Neyman-Pearsons lemma, asymptotic theory for Likelihood-Ratio tests, locally most powerful tests, and tests based on the ML-estimator. Bayesian tests, which may differ dramatically from their classical counterparts. Permutation tests and conditional tests. Robustness. Sequential tests. Tests in time series analysis.

Confidence intervals: From tests to intervals. The pivot method. Bayesian intervals.

Literature
The main textbook will be :  J.K.  Lindsay, Parametric Statistical Inference, Oxford University Press, 1996. ISBN 019 852359 9.
We will, however, also use some other material.