Welcome, Guest
Please Login or Register.    Lost Password?
Monolix Forum Statistical Model Statistical Model
Residual error models (1 viewing) (1) Guest
Go to bottom Favoured: 1
TOPIC: Residual error models
#86
Residual error models 9 Months, 2 Weeks ago  
HPLC results are typically heteroscedastic. Therefore I tend to using proportional residual error models. With Monolix these appear to produce worse fits as a rule than do additive error models. Is this just by chance?
Ingolf ()
User Offline Click here to see the profile of this user
Logged Logged  
 
The administrator has disabled public write access.  
#88
Re:Residual error models 9 Months, 1 Week ago  
Very small observed values in the data are difficult to fit with a purely proportional error model. A combined error model ( y = f + (a+b*f)e ) is preferred. You can estimate the constant term a, or fix it to any small value.

Marc
Marc Lavielle ()
User Offline Click here to see the profile of this user
Logged Logged  
 
The administrator has disabled public write access.  
Go to top
Powered by FireBoardget the latest posts directly to your desktop