DOC: Improve docs of regression_diagnostics.html, stats.html, summary

[luke396]

regression diagnostics:

• Add explanations for Durbin-Watson and Kurtosis.
• Include DW-test and Breusch-Godfrey test for autocorrelation.
• Reorder some parts for better readability

stats.html

• Reorder and add subtitle

summary

• Add notes section link to regression diagnostics

closes #9099

[luke396]

Hi @josef-pkt, could you find some time to review this PR? I'll make any necessary updates. Following this PR, I'll continue working on what we discussed in issue #9099.

[josef-pkt]

Thanks, looks good

can you squash this, then I will merge it

(I'm taking a break from robust and try to catch up with other PRs)

[luke396]

Great work on robust!

Do you mean we should address all the issues mentioned in #9099 and merge them at once? If so, I'll continue working on this and push to current PR. (My original plan was to fix them in separate PRs, but if you prefer to address them in a single PR, I'll start working on that.)

[josef-pkt]

A bit larger PRs, i.e. medium small PRs, are better for reviewing, combining related parts.
Small PRs still need attention so I still need to get my mind unstuck from whatever I'm doing unless they are bug-fixes limited to a relatively narrow piece of code.
Large PRs can take a large amount of time to review and so might also get delayed.

[luke396]

Thanks for the explanation!

I understand the importance of attention and time. Having a properly sized PR is indeed what we want.

[luke396]

I attempt to classify the methods, but I am uncertain about their academic accuracy.

[luke396]

I attempted to add the reference using a Sphinx hyperlink, but was unsuccessful as we lack a corresponding file (e.g., .rst, .md) instead of the .ipynb format.

Maybe there is a better way to finish.

[luke396]

Hello, @josef-pkt. I believe this pull request is ready for your review and will close the corresponding issue once approved.

Please take a look and feel free to leave any comments at your earliest convenience.