TMA4315: Generalized linear (mixed) models, autumn 2025

Exam Help

The exam is on 1st December, starting at 3pm.

The regulations say "Specified printed and hand-written support material is allowed. A specific basic calculator is allowed." This means a calculator and an A4 sheet: it does not need to be stamped or yellow (we used to have this rule but no longer). You can write on both sides. As far as I'm aware, you can print out something hand written on a tablet. You can also bring maths/stats tables, specifically "Tabeller og formler i statistikk (Tapir forlag, Fagbokforlaget)" (which you should be able to get from the bookshop in the university).

The exam will be written in English, but you can answer in English or Norwegian. This is not a language exam, so Bad English is acceptable. If in doubt, keep your sentences short and simple. Previous exams can be found here, and exam questions are on each module, which will also be helpful.

There will be two revision sessions:

* in Kjelhuset: KJL23 on Wednesday 26th November 13:00-15:00.

* in R4 (behind the cafe in Realfagbygget) on Friday 28th November 12:00-14:00.

I will be there for the whole time, so turn up when you can. If you can't make it at that time, I'm happy to get emails, or you can meet me in my office.

Good luck!

Timetable

The plenary lecture and exercise session will be on Monday, and the interactive lecture will be on Wednesday. If there are timetable clashes, we can move these around.

  • Lecture: Monday 10:15 - 12:00 in R4
  • Exercise session: Monday 12:15 - 14:00 in R4
  • Interactive Lecture: Wednesday 12:15 - 14:00 in B4

Links to all of the material will be put on this page, and the material itself is on GitHub All materials will be found on GitHub at this web link (the source code is here). There is some hidden material below, that will be revealed as we go through the modules: this already links to the module files.

We will be using Fahrmeir et al. as a textbook (this should be freely available for NTNU students. The link goes to the second edition, so some section numbers might be wrong. Tell me if you find an error!).

Module 1

Module 2

This module is split over 2 weeks, and also has a compulsory exercise, where you can make an R package to do what the lm() function already does. If you can do this, it means you understand how regression works (as well as how to make an R package). The deadline for handing in the compulsory exercise will be the end of Friday 12th September.

The Compulsory Exercise is here. It should be done in small groups (where "small" means about 1-5). Collaboration between groups is also fine: this isn't a competition. Whilst LLMs like ChatGPT aren't banned, just asking an LLM to do the whole exercise isn't a good way to learn, so I would suggest using it as a last resort when you're stuck. If you do use an LLM, please indicate where. You won't lose marks for asking for help from an LLM, just as you won't for asking for help from a friend. Getting an LLM or a friend to do your work for you, though… For what it's worth, in the "real world" the problem with LLMs is that you have to be confident they are correct, so you should understand their output.

The Notes contain all of the information in the Module, for both weeks.

There are also a couple of mathematical results from Härdle & Simes These are the Week 1 files, the presentation and interactive work:

These are the Week 2 files (some are still being editted):

Module 3

This module is split over 2 weeks. The Notes contain all of the information in the Module, for both weeks.

These are the Week 1 files, the presentation and interactive work:

These are the Week 2 files, the presentation and interactive work:

Here are a couple of videos. They should be embedded in the notes, but in case there is a problem, here are the links.

Module 4

This module is split over 2 weeks.

The Notes contain all of the information in the Module, for both weeks.

These are the Week 1 files, the presentation and interactive work:

These are the Week 2 files, the presentation and interactive work:

Module 5

There is a compulsory exercise associated with this week. the deadline is the end of Friday 17th October. This module is only one week. The Notes contain all of the information in the Module.

These are the presentation and interactive work:

I also made some videos. These should be embedded in the notes, but they are also here:

Module 6

This module is only one week.

The Notes contain all of the information in the Module.

These are the presentation and interactive work:

Module 7

This module is two weeks long.

The Notes contain all of the information in the Module.

These are the presentation and interactive work for Week 1:

And these are the presentation and interactive work for Week 2:

Module 8

The final exercise is here. The deadline is midnight on Friday 14th November (and not October 8th - my apologies if you panicked over that). You should do this in groups, and hand in on Backboard. You need to join a group on Blackboard, for compulsory exercise 3 (NOT compulsory exercise 1 or 2!), and from there you can submit your assignments, as many times as you want. As always, tell me if this isn't working, and I can help.

The Notes contain all of the information in the Module.

These are the presentation and interactive work:

Module 9

This is a summary of the course. There is no interactive material, but we will have a session if you want help with the exercise or revision for the exam.

2025-11-24, Bob O'Hara