Modify priors on `eta_sd` and `inf_feedback`

[kaitejohnson]

This PR addresses #227. It modifies the stan model to take in a mean of the RW step size in R(t), eta_sd, and then it tests it using the subset of forecast dates and locations set up in https://github.com/CDCgov/wastewater-informed-covid-forecasting/blob/prod/_targets_subset_benchmarking.R

Benchmarking results can be found here in the files labeled as subsets https://github.com/CDCgov/wastewater-informed-covid-forecasting/tree/205-test-out-less-informative-eta-sd/output/benchmarking. In the overall plots, the things to pay attention to are:

• pink (LHS) is this current proposal
• purple (RHS) is v1.0.0 (the currently, too strong infection feedback prior and two weak prior on eta_sd)

When we ran this on all forecast dates and location with a slightly more informative prior on eta_sd, the overall scores did look improved, though here the overall scores do not overall, if you look at the one by forecast date you see this is largely driven by 2024-01-08 as the pink (this proposed change) is an improvement most other times.

[github-actions[bot]]

Thank you for your contribution @kaitejohnson :rocket:! Your github-pages is ready for download :point_right: here :point_left:! _{(The artifact expires on 2024-11-19T19:10:48Z. You can re-generate it by re-running the workflow here.)}

[seabbs]

No objection. The current proposal seems reasonable. We should talk about identifiability in the manuscript

[kaitejohnson]

Updated changelog