Messages
Good luck on the exam on Monday! Please note that unlike last year, you are not allowed to bring your own sheet of notes into the exam. You may still bring a calculator, but you will not be needing one.
We are approaching the final week of lectures! On Tuesday, I will go through last year's resit exam (in Norwegian: kontinuasjonseksamen), which you can find here on the semester webpage. On Wednesday, I will give a summary of the main parts of the syllabus, and there will be time for questions.
Next week, we will go through last year's exam and problem 3 from last year's mandatory assignment. Both are available from last year's webpage.
Please keep in mind that last year's exam paper was too difficult, so don't panic if you struggle with it. I recommend you try all the problems in your own time (so don't do it as a timed paper), and maybe read through the solutions before the class.
In the class on Tuesday we will go through exercises 42, 44, 45, 46, 47, 48, 53(c) and 62. This will finish off the non-coding exercises in the course.
In the lecture on Wednesday we will finish chapter 8, looking at convergence diagnostics for MCMC.
On Tuesday, we will go through solutions to the Oblig, along with exercises 30, 31, 36, 37 and 38.
On Wednesday, we will look at the Metropolis-Hastings algorithm and Gibbs sampling, covering everything from page 53 to 59.
There will be another drop-in session today where you can ask questions about the oblig (or anything else from the course).
A correction to Problem 2 in the oblig has been made. In line 4 of the algorithm given in part (e), it now correctly says:
α ← exp{?(θ') - ?(θs)}
The pmf in Problem 1 in the oblig has now been corrected.
The mandatory assignment (oblig) will be made available today, Thursday, at noon. You will have 14 days to complete the assignment.
Next week there will be a drop-in session on Tuesday, where you can ask questions about the oblig (or anything else from the course). On Wednesday, we will cover Chapters 6 and 7.
For the class on Tuesday, please do exercises 24-29.
On Wednesday, we will aim to cover Section 5.2, finishing on page 44.
For the class on Tuesday, please do exercises 22 and 23.
I will be away on Wednesday, so Per August Moen will be your substitute teacher. He will continue onwards from 35 and aim to finish Section 5.1.
For the class on Tuesday, please do exercises 16-21. I mistakenly thought I had given you exercise 16 last week, so I covered in class already. I will briefly go over it again in the class next week as well so that we are all on the same page.
On Wednesday, we will begin on Chapter 5, covering all the material up to approximately page 34.
For the class on Tuesday, please do exercises 12-15.
On Wednesday, we will cover Chapter 3 and 4, finishing on page 29 (please note that Section 4.2 is not on the syllabus).
For the class on Tuesday, please do exercises 10 and 11.
On Wednesday, we will finish Chapter 2 in the notes (page 24).
Next week we will have a lecture on Tuesday and a class on Wednesday (rather than the usual formula with classes on Tuesdays and lectures on Wednesdays). In the lecture on Tuesday, we will continue out studies of the multivariate Gaussian distribution, finishing on approximately page 17 in the lecture notes.
For the class on Wednesday, please do exercises 6-9. I will be away on Wednesday, so Professor Nils Lid Hjort will be your substitute teacher.
The code for exercise 3 is available here on the course webpage (you may have to right click and click on "save link as" for it to download correctly).
Also, the bonus Python introduction video is now available here: https://youtu.be/nGh-qNHzqXQ
For the class on Tuesday, please do exercises 1 - 5.
In the lecture on Wednesday, we will see more examples of conjugate priors, and start looking at the multivariate Gaussian distribution (which will stay with us for quite some time). This will take us to approximately page 14.
Welcome to STK4021 - Applied Bayesian Analysis! I look forward to meeting you next week for the first set of lectures.
The syllabus for the course is collected in a compendium of lecture notes which can be found here on the course webpage. These include all the lecture material as well as the accompanying exercises, which I highly encourage you to attempt in your own time before seeing the solutions presented in classes. The classes will also go through solutions to the coding exercises, of which there are many scattered throughout the notes.
If you have any questions at all about the course please reach out to me via email at dchristensen@nr.no.
See you on Tuesday, August 20, at Auditorium 2 in Vilhelm Bjerknes' hus!