FMS091/MASM11: Monte Carlo and Empirical Methods for Stochastic Inference
Current information for spring 2017.
The course has now changed course code to fmsn50/masm11. The new home page can be found HERE
Basic course in mathematical statistics and at least one of
Stationary Stochastic Processes (FMS045/MAS210) or
Markov processes (FMS180/MAS204)
- 4 hours lectures each week (total 26 h)
- 2 hours computer assignments each week (total 12 h)
Simulation-based methods of statistical analysis. Markov chain Monte Carlo methods for complex problems, e.g. Gibbs sampling and the Metropolis-Hastings algorithm. Bayesian modeling and inference. The resampling principle, both non-parametric and parametric. Methods for constructing confidence intervals using resampling. Resampling in regression. Permutations test as an alternative to both asymptotic parametric tests and to full resampling. Examples of more complicated situations. Effective numerical calculations in resampling. The EM-algorithm for estimation in partially observed models.
The course aim (in Swedish).
Examination consists of home assignments which will be
handed out during the course.
Geof H. Givens and Jennifer A. Hoeting Computational Statistics
Second Edition (2012) and some
additional handouts. The book is available from KF-sigma.
Course administrator and lecturer (VT17)
Magnus Wiktorsson, MH:130
phone: 046-222 86 25