Dear all,
I'm using SAEM algorithm with option ETASAMPLES=1 ISAMPLE=10 EONLY=1 to draw
random eta samples from individual posteriors, according to the NONMEM user
guide. However, I realized that ISAMPLE can not be larger than 10 otherwise
NONMEM will generate only two samples per individual as if ISAMPLE=2. I was
wondering if there's way if I wish to have for instance 50 ISAMPLES?
Additionally, in the example from the NONMEM user guide of using ETASAMPLES=1
EONLY=1, NBURN is set to 200. From my understanding, for such purpose, SAEM
uses parameter estimates from previous $EST step and only draws random samples
around them and there's no maximization step. I guess NBURN=1 would be
sufficient?
Any comment is appreciated!
Warm regards,
Tingjie Guo
ISAMPLE in SAEM algorithm with ETASAMPLES=1
3 messages
2 people
Latest: Oct 08, 2020
Tingjie:
The ETASAMPLES method is better developed in NONMEM 7.5.0, along with the
description of it in nm750.pdf
Maximal ISAMPLE can be increased with
$SIZES ISAMPLEMAX=250
for example.
Robert J. Bauer, Ph.D.
Senior Director
Pharmacometrics R&D
ICON Early Phase
820 W. Diamond Avenue
Suite 100
Gaithersburg, MD 20878
Office: (215) 616-6428
Mobile: (925) 286-0769
[email protected]<mailto:[email protected]>
http://www.iconplc.com
Dear Robert,
Thanks! Do you also agree with NBURN=1?
Warm regards,
Tingjie
Quoted reply history
On Thu, Oct 8, 2020, at 06:33, Bauer, Robert wrote:
> Tingjie:
> The ETASAMPLES method is better developed in NONMEM 7.5.0, along with the
> description of it in nm750.pdf
> Maximal ISAMPLE can be increased with
>
> $SIZES ISAMPLEMAX=250
>
> for example.
>
> Robert J. Bauer, Ph.D.
> Senior Director
> Pharmacometrics R&D
> ICON Early Phase
> 820 W. Diamond Avenue
> Suite 100
> Gaithersburg, MD 20878
> Office: (215) 616-6428
> Mobile: (925) 286-0769
> [email protected]
> www.iconplc.com
>
>
>
>