|
Abstract:
|
Bayesian statistical methods are widely used in many science and engineering areas including machine intelligence, expert systems, medical imaging, pattern recognition, decision theory, data compression and coding, estimation and prediction, bioinformatics, and data mining.
These course notes present the basic principles of Bayesian statistics. The first sections explain how to estimate parameters for simple standard statistical models (normal, binomial, Poisson, exponential), using both analytical formulas and the free WinBUGS data modelling software. This software is then used to explore multivariate hierarchical problems that arise in real applications. Advanced topics include decision theory, missing data, change point detection, model selection, and MCMC computational algorithms.
Students are assumed to have knowledge of basic probability. A standard introductory course in statistics is useful but not necessary. Additional course materials (exercises, recorded lectures, model exams) are available at http://math.tut.fi/~piche/bayes |