How to implement multiple error model?

Hallo,

I have been working on population and covariate projects using PBPK model for a while,

I just wonder, I have several data set (from organ and serum, I am working on nuclear medicine) and I would like to fit these data simultaneously, is it possible to set different error model to each data and estimate residual error from each data in Simbiology?

and I just thinking if this doesnt work, maybe I can just fix different residual error for each data (I derived this residual error from another source) and I am not fitting it during the fitting, is it possible to do this?

Thank you very much for your answer, looking forward to it.

Answer 1

[Arthur Goldsipe]

It depends on whether you are doing nonlinear regression or nonlinear mixed-effects modeling. If you are doing nonlinear regression, you can use the ErrorModel option of sbiofit to specify a different error model for each response that you are fitting. For example, setting ErrorModel to {'constant', 'proportional'} will use a constant error model for the first response and a proportional error model for the second response.

Unfortunately, we don't yet the same capabilities in sbiofitmixed, if you are doing nonlinear mixed-effects modeling.

I hope that answers your question.

-Arthur

[Deni Hardiansyah]

Arthur,

Thank you very much for your answers,

I am working on covariate model selection with mixed-effect modeling and many types of data, it would be nice if I can fit individual error model for each data, I hope Simbiology can have this in the future.

Have a wonderful day!!

Deni

[Arthur Goldsipe]

Deni, I've relayed your request to the Statistics and Machine Learning team, since they are the currently maintaining the code for nonlinear mixed-effects modeling.

[Deni Hardiansyah]

Thanks!!