Wikipedia:Peer review/Bayesian model reduction/archive1

Bayesian model reduction

I've listed this article for peer review because this is my first contribution and any feedback is welcome.

Thanks, Peter. Peterzlondon (talk) 19:56, 18 May 2018 (UTC)

@Peterzlondon: I'm new to Wikipedia and I know nothing about statistics, so don't expect much help from me. However, here are a few questions regarding references:

• The article begins with "Bayesian model reduction [1][2]". Here, the references indicate that's the method's name, right?
• In the neurobiology subsection of the applications section, you use reference [5] mid-sentence. Is there any reason for it?
• Is there any reason you "define" the references the first time you use them, instead of down in the reference list? Given the amount of information you include in them, they take up a lot of space.
• Same question about not using template:r, which I personally like.

Additional questions:

• Is there any non-aesthetic reason you number the equations? I don't see it much here in Wikipedia (I'm not saying it's a bad thing!).
• Do you think the article could fit in any of the subcategories of category:statistical methods? I'd just add it to that category, but articles should always be in the narrowest subcategory possible.
• Finally, and most importantly: What exactly is the article's audience? Is it beginner statistics students, beginner statisticians, or somewhere in between?

I'll do the changes I want once you answer.

Professor Proof (talk) 22:12, 29 May 2018 (UTC)

@Professor Proof: Thanks for going through this! I'll reply to each point in turn:

• Regarding the references in the first sentence - "Bayesian model reduction [1][2]". The statistical method was first introduced in reference 1, for a specific application. Reference 2 generalised it to broader applications and gave it the name 'Bayesian model reduction', which is the title of the article. So I thought it better to include both references, to ensure they each get equal credit. However, it would probably be OK just to cite reference 2 here, given reference 1 is detailed below.
• In the sentence 'it has been proposed [5] that the brain...' I used the reference mid-sentence to make it clear that the reference is linked to the proposal. At the end of the sentence, I reference another paper (about sleep) and I didn't want to cause confusion. If you think this is inelegant, please feel free to adjust.
• Regarding 'defining' the references the first time I use them - I am not sure I fully understand, please can you give an example?
• Thank you for telling me about template template:r - I wasn't aware of that. Happy to use it.

Additional questions:

• I numbered the equations to make it easier to reference (I could say 'take a look at Equation 4 on the Wikipedia page'). However, it does introduce an annoyance - the numbering is not compatible with Wikipedia's graphical equation editor. So this may be good reason for getting rid of them.
• The page is listed in the more specific category:Bayesian_statistics
• The audience I had in mind is anyone working in science or engineering, who use mathematical models to understand their measurements. They could be students, academic researchers or people working in industry. I expected the reader to have an awareness of Bayesian statistics and an interest in Scientific modelling.

Thanks again.

Peterzlondon (talk) 09:33, 30 May 2018 (UTC)

@Peterzlondon: OK, made the change I wanted to make (moving refs to the end of sentences and "defining them at the bottom"). What do you think?

As for the categories, category:Bayesian statistics isn't "more detailed" (i.e. a subcategory, sub-subcategory, etc.) than category:statistical methods. It makes sense to me—some subjects in Bayesian statistics aren't methods.

I wish I could help more, but I don't think I have the skills to really improve things. It's great for a first article, by the way. The subject is a bit heavy for me, but it seems well-presented and well-organized. Professor Proof (talk) 18:38, 30 May 2018 (UTC)