From: Mark Sale <msale@medlib.iaims.georgetown.edu>
Subject: OMEGA and SIGMA in INFN
Date: 23 Jul 1997 16:07:51 -0400
Dear users,
Has anyone every addressed OMEGA and SIGMA in the INFN subroutine? THETA is an argument to this routine, but OMEGA and SIGMA are not.
Mark Sale MD
Assistant Professor of Medicine
Georgetown University
msale@medlib.georgetown.edu
202-687-8242
FAX 202-687-6437
OMEGA and SIGMA in INFN
From: Mats Karlsson <mats.karlsson@biof.uu.se>
Subject: Re: OMEGA and SIGMA in INFN
Date: 23 Jul 1997 17:16:23 -0400
Dear Mark,
Here is an INFN-routine I used quite some time ago to summarize the result from a 100 subproblem NONMEM run. Unfortunately, it is not commented and the coding is "ugly" but maybe of some use still.
Best regards,
Mats Karlsson
subroutine infn (icall,theta,datrec,indxs)
dimension theta(*),datrec(*),indxs(*)
dimension s(100),ss(100),sss(100)
dimension the1(100),the2(100),the3(100),om4(100)
dimension om1(100),om2(100),om3(100),si1(100)
dimension sr(100),ssr(100),e(100),sd(100)
double precision theta,object,s,ss,sss
double precision thetaf,omegaf,sigmaf
double precision the1,the2,the3
double precision om1,om2,om3,om4,si1
double precision sr,ssr
integer nprob,iprob,un,un2,un3
DATA un /80/
DATA un2 /90/
DATA un3 /95/
common /rocm6/ thetaf(20),omegaf(10,10),sigmaf(10,10)
common /rocm8/ object
common /rocm9/ iere,ierc
common /rocm10/ nrep,irep
if (icall.eq.0) then
nprob=100
ne=0
iprob=0
do 3 i=1,8
ssr(i)=0
3 sr(i)=0
do 5 i=1,nprob
s(i)=0
ss(i)=0
sss(i)=0
the1(i)=0
the2(i)=0
the3(i)=0
om4(i)=0
om1(i)=0
om2(i)=0
om3(i)=0
5 si1(i)=0
endif
if (icall.eq.1) then
iprob=iprob+1
endif
if (icall.eq.3) then
s(iprob)=object
ss(iprob)=iere
sss(iprob)=iprob
the1(iprob)=thetaf(1)
the2(iprob)=thetaf(2)
the3(iprob)=thetaf(3)
om1(iprob)=omegaf(1,1)
om2(iprob)=omegaf(2,2)
om3(iprob)=omegaf(3,3)
om4(iprob)=omegaf(4,4)
si1(iprob)=sigmaf(1,1)
if(iprob.eq.nprob) then
do 6 i=1,nprob
6 write(un,*)s(i),ss(i),sss(i)
if (iprob.eq.nprob) then
do 7 i=1,nprob
7 write(un2,8)the1(i),the2(i),the3(i),
1om1(i),om2(i),om3(i),om4(i),si1(i),sss(i)
8 FORMAT(9F9.4X)
endif
endif
if (iere.eq.0) then
ne=ne+1
sr(1)=sr(1)+the1(iprob)
ssr(1)=ssr(1)+the1(iprob)**2
sr(2)=sr(2)+the2(iprob)
ssr(2)=ssr(2)+the2(iprob)**2
sr(3)=sr(3)+the3(iprob)
ssr(3)=ssr(3)+the3(iprob)**2
sr(4)=sr(4)+om4(iprob)
ssr(4)=ssr(4)+om4(iprob)**2
sr(5)=sr(5)+om1(iprob)
ssr(5)=ssr(5)+om1(iprob)**2
sr(6)=sr(6)+om2(iprob)
ssr(6)=ssr(6)+om2(iprob)**2
sr(7)=sr(7)+om3(iprob)
ssr(7)=ssr(7)+om3(iprob)**2
sr(8)=sr(8)+si1(iprob)
ssr(8)=ssr(8)+si1(iprob)**2
endif
if (iprob.eq.nprob) then
if (ne.gt.0) then
e(1)=sr(1)/ne
e(2)=sr(2)/ne
e(3)=sr(3)/ne
e(4)=sr(4)/ne
e(5)=sr(5)/ne
e(6)=sr(6)/ne
e(7)=sr(7)/ne
e(8)=sr(8)/ne
if (ne.gt.1) then
sd(1)=sqrt((ssr(1)-sr(1)**2/ne)/(ne-1))
sd(2)=sqrt((ssr(2)-sr(2)**2/ne)/(ne-1))
sd(3)=sqrt((ssr(3)-sr(3)**2/ne)/(ne-1))
sd(4)=sqrt((ssr(4)-sr(4)**2/ne)/(ne-1))
sd(5)=sqrt((ssr(5)-sr(5)**2/ne)/(ne-1))
sd(6)=sqrt((ssr(6)-sr(6)**2/ne)/(ne-1))
sd(7)=sqrt((ssr(7)-sr(7)**2/ne)/(ne-1))
sd(8)=sqrt((ssr(8)-sr(8)**2/ne)/(ne-1))
endif
endif
do 9 i=1,8
9 write(un3,*)ne,e(i),sd(i)
endif
endif
end
From: "Gibiansky, Ekaterina" <Katyag@mocr.oapi.com>
Subject: RE: OMEGA and SIGMA in INFN
Date: 23 Jul 1997 17:18:30 -0400
Mark,
OMEGA and SIGMA are in the following COMMON BLOCK:
COMMON /ROCM6/ THETAF(20), OMEGAF(10,10), SIGMAF(10,10)
I have a small INFN that outputs final THETAs, OMEGA, SIGMA, individual ETAs and standard errors into different files. I can e-mail it to you if you are interested.
Katya Gibiansky
*************************************************
Otsuka America Pharmaceuticals, Inc.
2440 Research Blvd, Rockville,MD 20850
E-mail: katyag@mocr.oapi.com
Phone: (301)-527-4911