STK9190 – Bayesian nonparametrics
Course description
Schedule, syllabus and examination date
Course content
Statistical analysis involves first setting up a model for data in terms of certain unknown parameters. Bayesian analysis proceeds by placing a prior distribution on these parameters and then deriving and using relevant aspects of the consequent posterior distribution. Bayesian nonparametrics is the extended branch of such modelling and analyses where the parameter of the model is of very high or infinite dimension, as when one models an unknown density, regression, or link function. This calls for more complex mathematics and computational schemes than for the classical cases where the parameter is of low dimension. There are links to and implications for machine learning.
Learning outcome
After having completed the course you will have learned some of the more prominent nonparametric prior constructions and ensuing posterior calcuations:
- the Dirichlet process
- the Beta process
- Gaussian processes
- bigger hierarchical models
- applications with real data
After having completed the course you will also be able to:
- present, on a scientific level, a short thesis on a chosen topic of relevance, selected in collaboration with the lecturer
Admission
PhD candidates from the University of Oslo should apply for classes and register for examinations through Studentweb.
If a course has limited intake capacity, priority will be given to PhD candidates who follow an individual education plan where this particular course is included. Some national researchers’ schools may have specific rules for ranking applicants for courses with limited intake capacity.
PhD candidates who have been admitted to another higher education institution must apply for a position as a visiting student within a given deadline.
Prerequisites
Recommended previous knowledge
STK1100 – Probability and Statistical Modelling
STK1110 – Statistical Methods and Data Analysis
STK4021 – Applied Bayesian Analysis/STK9021 – Applied Bayesian Analysis
and one of the following courses:
STK2100 – Machine Learning and Statistical Methods for Prediction and Classification
STK2120 – Statistical Methods and Data Analysis 2 (discontinued)
STK3100 – Introduction to Generalized Linear Models
Overlapping courses
10 credits overlap with STK4190 – Bayesian nonparametrics
Teaching
3 hours of lectures/exercises per week.
Examination
Depending on the number of students, the exam will be in one of the following four forms:
1. Only written exam
2. Only oral exam
3. A project paper followed by a written exam.
4. A project paper followed by an oral exam/hearing.
For the latter two the project paper and the exam counts equally and the final grade is based on a general impression after the final exam. (The two parts of the exam will not be individually graded.)
The form of examination will be announced by the teaching staff by 15 October/15 March for the autumn semester and the spring semester respectively.
In addition, each PhD candidate is expected to give an oral presentation on a topic of relevance chosen in cooperation with the lecturer. The presentation has to be approved by the lecturer for the student to be admitted to the final exam.
Examination support material
No examination support material is allowed.
Language of examination
Subjects taught in English will only offer the exam paper in English.
You may write your examination paper in Norwegian, Swedish, Danish or English.
Grading scale
Grades are awarded on a pass/fail scale. Read more about the grading system.
Explanations and appeals
Resit an examination
This course offers both postponed and resit of examination. Read more:
Withdrawal from an examination
It is possible to take the exam up to 3 times. If you withdraw from the exam after the deadline or during the exam, this will be counted as an examination attempt.
Special examination arrangements
Application form, deadline and requirements for special examination arrangements.
Evaluation
The course is subject to continuous evaluation. At regular intervals we also ask students to participate in a more comprehensive evaluation.