Hi
I would like to ask to NONMEM community whether it is possible to re-use the
results of a nonmem model run when starting bootstrapping.
As I understand, a bootstrap run using PsN starts with a single calculation of
the model of interest and data set as is. Once this initial model run
has comp
Dear Jos,
If you haven't run the original model (indicated by absence of lst-file with
the same prefix) psn will run it before the bootstrap. If you have run it,
PsN will not rerun it. It will use the results in the corresponding lst
file. If you have run it and called it something else, you co
Dear Mats
My question is more related to boot strapping itself.
if data set and model I am working with takes long time to run, then during
bootstrap run I only run $EST and skip $COV step. We will get confidence
interval for parameter from bootstrap results.
What is your opinion on that?
Best
Hi Anuba,
I would agree with your suggestion. That is how I usually do it.
Best regards,
Mats
Mats Karlsson, PhD
Professor of Pharmacometrics
Dept of Pharmaceutical Biosciences
Faculty of Pharmacy
Uppsala University
Box 591
75124 Uppsala
Phone: +46 18 4714105
Fax + 46 18
Dear Anuhbia,
Since bootstrap is performed to replace the estimation of precision that
is performed in the $COV step, you definitively do not need the $COV
option when you do a bootstrap.
Then CI are obtained based on the quantiles of the bootstrap estimates.
Kind regards,
Pascal Girard, Ph
PhD/early stage research studentship available in a Marie Curie
EU-funded joint project by University of Warwick, AstraZeneca,
Fraunhofer-Chalmers Research Centre for Industrial Mathematics, and
Swedish University of Agricultural Sciences on Modelling Dose Response
Time Outcome Relationships.
