Lab 5 logit transformation on rho

[klomori]

In lab 5, logit_rho is entered as a parameter, but it is later logit inverse-transformed into rho (in the objective function section), where rho is used later in the script (not logit_rho). Why does rho need to be transformed from logit_rho instead of using rho as the estimated parameter directly?

[James-Thorson]

I recommend trying it both ways ;)

Short answer: some of the methods used only work when rho is between -1 and 1, so it prevents an error to bound it and is easier to do in TMB than the R side