search

karimn / covid-19-transmission

0 stars 0 forks source link

Parametric trend #12

Open karimn opened 4 years ago

karimn commented 4 years ago

Bottom of page 1 and top of page 2

heterogen_sub_comm.pdf

karimn commented 4 years ago

@wwiecek in the cr(t) function you proposed, the parameter f can only be 0.5 if you want cr(t) to start at 1. Is that intentional?

This is not a very easy functional form to make hierarchical.

wwiecek commented 4 years ago

Not intentional at all, I just can't be trusted even with the simplest tasks! Logistic should be scaled by 1-f, not 2f, so that it travels from 1 (no scaling) to f (minimum value) at a rate k, with some predefined mid-point.

We can have a pretty strong prior on k (the decline from 1 to f should be weeks, not months or years) and a more diffuse prior on f, which is in 0 to 1 range (so we could use a transform?)

Does that sound sensible? It is very much just an idea. If it sounds suspect I could check if there is literature for this.

karimn commented 4 years ago

Implement with hierarchy for the lambda or kappa parameters. Lambda no fully separated between regions.

wwiecek commented 4 years ago

LMK if I should have a look at something in particular, I'm focusing on the descriptive PDF

karimn commented 4 years ago

I think I'm set on this end of things. Please look at the imputed cases issue and let me know if we should do anything else with the priors. I think you might have found the heart of our convergence problems there.

On Wed, Jul 1, 2020 at 9:52 AM Witold Wiecek notifications@github.com wrote:

LMK if I should have a look at something in particular, I'm focusing on the descriptive PDF

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/karimn/covid-19-transmission/issues/12#issuecomment-652431504, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABB5543EY67DZ66MHSCAETRZM5RJANCNFSM4ODVLXNQ .

karimn commented 4 years ago

What's next: hierarchical structure.

wwiecek commented 4 years ago

What's next: hierarchical structure.

I think we will need a hierarchical structure, but if we have to prioritise maybe jsut starting with non-hierarchical paramters for each region (or maybe for each country? I'm not sure) and fixing hierarchical structure for R0's. What do you think?

karimn commented 4 years ago

Agreed. I just added this to remind us where we left off with this issue.

On Thu, Jul 9, 2020 at 6:04 AM Witold Wiecek notifications@github.com wrote:

What's next: hierarchical structure.

I think we will need a hierarchical structure, but if we have to prioritise maybe jsut starting with non-hierarchical paramters for each region (or maybe for each country? I'm not sure) and fixing hierarchical structure for R0's. What do you think?

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/karimn/covid-19-transmission/issues/12#issuecomment-656036014, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABB55ZRDOSP3PJQVBSVL6DR2WI4DANCNFSM4ODVLXNQ .