Probability for Inference or Linear Statistical Models easier?

[Prelud3]

Hi guys
I am struggling between MAST30020 Probability for Inference and MAST30025 Linear Statistical Models
Apparently I can only choose one of these as my elective.
Can anyone give me some suggestion on which subject shall I enroll?
Many thanks

[96n_n]

Both are hard subjects. I’ve had friends drop out of Probability for Inference in the first few weeks because it was difficult to keep up. From what I’ve heard from other students who continued the subject, it is extremely time consuming because you’re trying to wrap your head around concepts. They said that the whole 12 weeks of learning the subject was very difficult (which was also affected by the number of assignments the lecturer handed out), HOWEVER, the exam was very standard to past papers. So if you know how to do past papers, you would be able to pass.

Linear Statistical Models is very difficult in its own right too. Yao-ban is a great lecturer - teaches the content very well and at first, the information he provides in the lecture seems pretty easy to understand - you can read the slides and understand the basics. BUT, despite understanding the basics, I sometimes thought I never truly understood them because I struggled with the assignments and the final exam. I knew no-one who got full marks in the assignments or at least close to full marks. Most people had to work together to tackle the questions. And the exam was pretty much horrendous (my personal experience - some said it was okay, but most said it was horrible) He made it much more difficult and different from the other past exams, and so, evidently he had to scale the scores.

Probably best for someone who has taken both subjects to answer this question though

[stolenclay]

I've done both of these subjects and was also taught by the current lecturers for those subjects.

I think your choice depends on what you want to do afterwards.

If there's any chance that you want to do further study in probability theory then Probability for Inference is the one you should choose; it would be quite difficult learning the material in that subject without a dedicated lecturer, and you will need to understand the material for more advanced study, e.g. any study of stochastic processes past the Stochastic Modelling subject. The workload is colossal, but the material in the subject is abstract and really does warrant it.

On the other hand, Linear Statistical Models is absolutely vital for understanding the terminology and the way of thinking for statistical regression problems. This is not even taking into account the fact that linear regressions themselves are staple in most basic statistical analyses. The subject (at least when I was taught by Yaoban) is slow but careful in its approach, making it theoretical and a bit dry as a result, but this is pretty much the only stage in your statistical education where anyone would care to explain/prove to you these results. If you come across the same results later on then chances are it'll be assumed knowledge or glossed over. If you like to keep a watertight understanding of the "why" in maths then you want to take this.

Both subjects are important for different reasons. I imagine that, for most people, Linear Statistical Models is the more sensible choice, but without more information I can't say for sure what would be suitable for you.