Add vignette with case study using new model_compare()

[florence-bockting]

Fixes #401

Description

This PR adds a vignette that demonstrate a case study, guiding the user through the Bayesian model comparison workflow using the new predictive measure API and the new model_compare() function.

This vignette completes a set of vignettes for loo version 3, focussing:

• predictive measures (elpd, rmse, r2, rps, ...)
• predictive schemes (loo, kfold, test,...)
• [this one] workflow/case study
• model comparison with model_compare() (focus on API not workflow)

This PR uses AI assistance.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.92%. Comparing base (de81c46) to head (22a4981).

Additional details and impacted files

@@              Coverage Diff              @@
##           pred_measure     #404   +/-   ##
=============================================
  Coverage         91.92%   91.92%           
=============================================
  Files                38       38           
  Lines              5068     5068           
=============================================
  Hits               4659     4659           
  Misses              409      409           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 22a4981 is merged into pred_measure:

• ✔️loo_function: 1.65s -> 1.64s [-2%, +1.03%]
• ✔️loo_matrix: 1.59s -> 1.59s [-0.82%, +1.04%]
  Further explanation regarding interpretation and methodology can be found in the documentation.