@whelanh1 That's v interesting re forest. So, like virus spreading with regional outbreaks significantly displaced in time, but combined in overall data. I'm thinking Sweden, Brazil, USA (as a whole) etc.
Looks like you're using similar method fundamentally to @MLevitt_NP2013, 1/n
Looks like you're using similar method fundamentally to @MLevitt_NP2013, 1/n
@whelanh1 @MLevitt_NP2013 2/n in using Gompertz to home in over time on a pedicted final value. So I'd have thought it to be reasonable, though since the transmission & population (susceptibles) change in & out of lockdown in these cases I'm not sure the 2 functions need to share inflections or ...
@whelanh1 @MLevitt_NP2013 3/n other parameters necessarily (could be wrong). Could u get better fit by relaxing that requirement? Then your new section of forest analogy woild fit something akin to what I was saying about modelling lockdown as limiting the accessible population. Also wondering if....
@whelanh1 @MLevitt_NP2013 4/4 you can drop.firat Gompertz - since we already know the outcome - & concentrate on fitting only the current phase, then add the earlier deaths to forecast a grand total.
Probably too tired so I apologise if much of this is nonsense (!). Talking tomorrow is good. All the best
Probably too tired so I apologise if much of this is nonsense (!). Talking tomorrow is good. All the best
