search

athowes / multi-agyw

Spatio-temporal estimates of HIV risk group proportions for AGYW across 13 priority countries in sub-Saharan Africa
https://athowes.github.io/multi-agyw
MIT License
2 stars 2 forks source link

Add PC prior on AR(1) parameter #12

Closed athowes closed 3 years ago

athowes commented 3 years ago

See Haakon's website and this paper

athowes commented 3 years ago

I think it's this theta = list(prior = "pc.cor1", param = c(<u>, <alpha>)), see here https://inla.r-inla-download.org/r-inla.org/doc/prior/pc.cor1.pdf. Any way to think of what u and alpha should be?

athowes commented 3 years ago

pc.cor1 is for a base model of correlation 1 and pc.cor0 is for a base model of correlation 0. I think there is good reason for the base model to be 1 in this case, as each measurement is of (in some sense) the same thing, just shifted by time.

athowes commented 3 years ago

Looking at this search https://github.com/search?l=R&q=rho+%3D+%22pc.cor1%22&type=Code, what I have seems to be pretty standard