Course: Bayesian Statistics

Course type: programme-based elective
Lecturer: Assoc Prof Mihael Perman, Assoc Prof Jaka Smrekar
 

Study programme and level Study field Academic year Semester
Applied statistics, second level All modules 2nd 1st

For the timeline see Curriculum.

Prerequisites:

  • Enrolment into the relevant academic year.

Content (Syllabus outline):
 

  • Bayesian models with one and more parameters. Connection with standard statistical methods. Hierachical models. Testing of models and sensitivity analysis. Bayesian  design of experiment.
  • Bayesian approach to evidence synthesis of multiple surveys, power priors, analysis of dependence of synthesis analysis on previous surveys.
  • Introduction into regression analysis. Analysis of variance and covariance. Hypothesis testing via Bayes factor, complexity and fit. Posterior probabilities of hypotheses – models, and influence of priors on them, training sample.
  • More on posterior probabilities, estimating parameters, central credibility interval, the importance of conjugated distributions. Gibbs sampler, convergence of estimates, algorithm Metropolis-Hastings. Posterior simulations. Some other specific models of Bayesian anlysis

Objectives and competences:

Basic knowledge of Bayesian statistics is acquired.

Bayesian methods are of great importance in practice. Therefore, experts with practical knowledge will present their experience in class.

Intended learning outcomes:
Understanding of basic concepts of Bayesian statistics.

 

Contact

Main contact:
e-mail: info.stat (at) uni-lj.si

Contact for administrative questions (enrolment, technical questions):
Katarina Erjavec Drešar
University of Ljubljana, Faculty of electrical engineering, Tržaška cesta 25, 1000 Ljubljana.
room num.: AN012C-ŠTU
phone: 01 4768 209
e-mail: katarina.erjavec-dresar (at) fe.uni-lj.si