Open seabbs opened 2 years ago
this was a pragmatic change - feel free to propose a more correct version.
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?
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).