Hi everyone;
I am currently performing an analysis where I am starting with iterative
two-stage, then SAEM, and finishing off with importance sampling. Here is
the relevant code:
$ESTIMATION METH=ITS INTERACTION NSIG=3 SIGL=6 CTYPE=3 NOABORT
NITER0 FILE=RUN19.TTS PRINT
$ESTIMATION METH=SAEM INTERACTION NOABORT CTYPE=3 NBURN00
NITER0 FILE=RUN19.SAE PRINT
$ESTIMATION METH=IMP INTERACTION ISAMPLE00 NITER EONLY=1
FILE=RUN19.IMP
$COVARIANCE
$TABLE ID TIME AGE WT SEX CLR CL V2R V2 V3R V3 EWRES CWRES NPDE
CPRED PREDI CPREDI MDV ETA1 ETA2 ETA3 ETA6
FILE=RUN19.CSV FORMAT=,1PE7.4 ONEHEAD NOPRINT
I notice that I am not getting standard errors for any of these runs,
except after the iterative two-stage portion of the analysis, which are
not especially useful. Can anyone shed some light as to what I am leaving
out of my code? I have been through the documentation and can not find
what I am doing wrong, although it is possible (likely!) that I am missing
something.
Thanks,
Mike Fossler
GSK
Standard errors using SAEM
Hi everyone;
I am currently performing an analysis where I am starting with iterative
two-stage, then SAEM, and finishing off with importance sampling. Here is
the relevant code:
$ESTIMATION METH=ITS INTERACTION NSIG=3 SIGL=6 CTYPE=3 NOABORT
NITER=100 FILE=RUN19.TTS PRINT=15
$ESTIMATION METH=SAEM INTERACTION NOABORT CTYPE=3 NBURN=300
NITER=800 FILE=RUN19.SAE PRINT=15
$ESTIMATION METH=IMP INTERACTION ISAMPLE=1000 NITER=10 EONLY=1
FILE=RUN19.IMP
$COVARIANCE
$TABLE ID TIME AGE WT SEX CLR CL V2R V2 V3R V3 EWRES CWRES NPDE
CPRED PREDI CPREDI MDV ETA1 ETA2 ETA3 ETA6
FILE=RUN19.CSV FORMAT=,1PE7.4 ONEHEAD NOPRINT
I notice that I am not getting standard errors for any of these runs,
except after the iterative two-stage portion of the analysis, which are
not especially useful. Can anyone shed some light as to what I am leaving
out of my code? I have been through the documentation and can not find
what I am doing wrong, although it is possible (likely!) that I am missing
something.
Thanks,
Mike Fossler
GSK
Try
$COVARIANCE UNCONDITIONAL
This will give you a standard error from the METHOD=IMP EONLY=1 step
even if there is some positive indefiniteness. But, be cautious about
its quality when such a message occurs.
Robert J. Bauer, Ph.D.
Vice President, Pharmacometrics
ICON Development Solutions
Tel: (215) 616-6428
Mob: (925) 286-0769
Email: [email protected]
Web: www.icondevsolutions.com
Quoted reply history
________________________________
From: [email protected] [mailto:[email protected]]
On Behalf Of [email protected]
Sent: Wednesday, December 02, 2009 9:57 AM
To: [email protected]
Subject: [NMusers] Standard errors using SAEM
Hi everyone;
I am currently performing an analysis where I am starting with iterative
two-stage, then SAEM, and finishing off with importance sampling. Here
is the relevant code:
$ESTIMATION METH=ITS INTERACTION NSIG=3 SIGL=6 CTYPE=3 NOABORT
NITER=100 FILE=RUN19.TTS PRINT=15
$ESTIMATION METH=SAEM INTERACTION NOABORT CTYPE=3 NBURN=300
NITER=800 FILE=RUN19.SAE PRINT=15
$ESTIMATION METH=IMP INTERACTION ISAMPLE=1000 NITER=10 EONLY=1
FILE=RUN19.IMP
$COVARIANCE
$TABLE ID TIME AGE WT SEX CLR CL V2R V2 V3R V3 EWRES CWRES NPDE
CPRED PREDI CPREDI MDV ETA1 ETA2 ETA3 ETA6
FILE=RUN19.CSV FORMAT=,1PE7.4 ONEHEAD NOPRINT
I notice that I am not getting standard errors for any of these runs,
except after the iterative two-stage portion of the analysis, which are
not especially useful. Can anyone shed some light as to what I am
leaving out of my code? I have been through the documentation and can
not find what I am doing wrong, although it is possible (likely!) that I
am missing something.
Thanks,
Mike Fossler
GSK