Open mattfidler opened 4 years ago
Variable | Meaning |
---|---|
OPRED | Original Individual predictions of model (ie eps=0, eta=ebeEta) |
DETA# | Sensitivity of ETA with respect to PRED |
OETA# | Original ETA value |
D_EPSETA#_x | Sensitivity of ETA with respect to R2 |
nlmixr
sets err
to 1 for all situations and estimates theta based error parameters for these values. Hence these error values would still need to be estimated (and on the same scale).
In nlmixr a linearized model would be addivitive with fixed component fixed to 1.
Like PsN. Since sensitivites are calculated it may not be as hard as I originally thought (with the exception of keeping track of f/dur/ etc)