Course type: compulsory
Lecturer: Aleš Žiberna, Ph.D., Professor
 

For the timetable see Curriculum.

Prerequisites:

• Enrollment in study year.

Content (Syllabus outline):
Statistical simulations:

• Introduction
• Estimating bias and standard errors
• Testing and evaluating statistical methods and tests
• Design of simulation studies
• Analysis and presentation of results
• Parallel computing

Resampling methods:

• Boostrap (standard errors, bias, test and confidence intervals)
• Permutation tests (statistical tests)
• Jackknife (standard errors, bias, model validation)
• Cross-validation (model validation)

Missing values:

• Types and mechanisms of missing values
• Methods for dealing with missing values with emphasis on:
  • Multiple imputations
  • EM algorithm

Monte Carlo Monte Chain methods:

• Gibbs sampling
• Metropolis-Hastings algorithm

Additional topics (time permitting)    

Objectives and competences:
The aim of the course is to enable the students for learning, adapting and using computer intensive methods in statistics. After the course students should be able to use these methods for solving real statistical problems that cannot be solved analytically.

Intended learning outcomes:
Students learn selected computer intensive methods and understand basic principles of these methods. They are able to use them on problems disused in the class and to adapt and use them on similar problems.