2024-12-08 update : changing center of prior.

[cbernalz]

Regarding #259 (comment) .

[kaitejohnson]

@cbernalz Could you provide a comparison of the posterior estimates of the phi parameter with both priors from the fit to the exponential and the data generated from the exponential? Ideally a density plot or histogram with the priors and posteriors overlaid

I am curious if with either of the prior values, the posterior estimate of the true phi is identifiable -- since this is generated with simulated data we know this.

ww-inference-model/vignettes/spatial_wwinference.Rmd

Line 246 in bc89b7b

phi = 0.4,

After doing that check, I would also consider increasing phi in the generation of the simulated data above, this way the strength of the correlation as a function of distance is increased in your simulation analysis

[cbernalz]

@cbernalz Could you provide a comparison of the posterior estimates of the phi parameter with both priors from the fit to the exponential and the data generated from the exponential? Ideally a density plot or histogram with the priors and posteriors overlaid

I am curious if with either of the prior values, the posterior estimate of the true phi is identifiable -- since this is generated with simulated data we know this.

ww-inference-model/vignettes/spatial_wwinference.Rmd

Line 246 in bc89b7b

phi = 0.4,

After doing that check, I would also consider increasing phi in the generation of the simulated data above, this way the strength of the correlation as a function of distance is increased in your simulation analysis

@kaitejohnson Here is the plot that results from the prior and posterior for $\phi$. The old prior is centered at 0.25 with sd 0.75. The new prior is centered at 1 with sd 0.4. The actual value is 1.5.

[kaitejohnson]

This looks reasonable to me! Model seems to struggle in both cases a bit with identifiability but this makes sense -- between the R(t) variability and lab-site level observation noise there are a lot of drivers of variation in C(t).

[kaitejohnson]

LGTM!