NMAK15014U Gaussian Graphical Models

Volume 2015/2016
Education

MSc programme in Statistics
 

Content
  • The multivariate Gaussian distribution; definition and basic properties
  • The Wishart distribution; definition and basic properties
  • The Markov property on undirected graphs
  • Conditional independence in the multivariate Gaussian distribution
  • Gaussian graphical models: definition and basic likelihood theory
  • Decomposable Gaussian graphical models
  • Model determination in Gaussian graphical models
  • The graphical lasso
  • Wishart and inverse Wishart distributions on Gaussian graphical models
  • Bayesian model determination in Gaussian graphical models
Learning Outcome

Knowledge:

Basic knowledge of the topics covered

Skill:

  • discuss and understand issues of conditional independence in Gaussian graphical models
  • discuss and understand basic algorithms for estimation and model determination in Gaussian graphical models
  • Ability to analyse simple multivariate datasets using Gaussian graphical models using standard software packages

 

Competences:

  • discuss and understand the role of decomposability for Gaussian graphical models
  • discuss and understand the role, similarities, and differences between model determination methods in Gaussian graphical models

 

Excerpts from suitable monographs and lecture notes.

Some basic understanding of statistical concepts and statistical distribution theory, including conditional distributions and conditioning eg Stat 1+2 plus Beting
6 weeks with 4 hours of lectures and 2 hours of exercises
1 week with 4 hours of lectures and 4 hours of practical exercises
  • Category
  • Hours
  • Exam
  • 27
  • Lectures
  • 28
  • Practical exercises
  • 4
  • Preparation
  • 135
  • Theory exercises
  • 12
  • Total
  • 206
Credit
7,5 ECTS
Type of assessment
Written assignment, 27 hours
Written homework.
Exam registration requirements

To register for the exam, the student has to present exercises in class

Aid
All aids allowed
Marking scale
7-point grading scale
Censorship form
No external censorship
One internal examiner
Re-exam

As ordinary exam. If the student has not presented the required exercises in class in order to register for the exam, he/ she can fulfil the requirements by presenting an exercise to the professor at the latest two weeks prior to the beginning of the re-exam week.

Criteria for exam assesment

The student must in a satisfactory way demonstrate that he/she has mastered the learning outcome of the course.