search

SACEMA / omicronSA

4 stars 0 forks source link

Changing logmean incubation period #5

Open seabbs opened 2 years ago

seabbs commented 2 years ago

Here

https://github.com/SACEMA/omicronSA/blob/49ef2d68125394cbb80f9fe2d7fe7681a90673f5/est_rt_ratios.R#L48

I see a change to the geometric mean. This is fine but if the idea is to change the mean of the distriubtion by some aamount (in the code 0.5) this is not correct as the mean and standard deviation of the lognormal both alter the natural scale mean (so need to use a transform to map vs a simple log addition).

pearsonca commented 2 years ago

this was a pragmatic change - feel free to propose a more correct version.

seabbs commented 2 years ago

EpiNow2 has built in conversion functions to map natural scale mean and sd to log scale. Could make use of those and pick a shorter incubation period explicitly?