R help - Jul 2009 - HELP: BRUGS/WinBUGS/RBUGS Response is a combination of random variables

Hi,

Is there anyone know if BUGS language allows the combination of variables as
response

such as
Y[i] <- a*X1[i]+b*X2[i]
Y[i] ~ dnorm(c,d)


It seems doesn't work in my model. The problem is between two ######.
The error message is 
>  modelCheck("BayesBioMarker.BUGS")
model is syntactically correct
>  modelData(paste("BUGS_data.txt",sep=""))
data loaded
>  modelCompile(numChains=1)
multiple definitions of node bm[1]


================================== model
            {
               for(iter in 1:numSubj){
                 bmd[iter] ~ dnorm(u[trt[iter]+1],sdbmd.precision);
                 ###############
                 bm[iter] <-  inprod(biomarker[iter,1:4], bta[1:4])
                 bm[iter] ~ dnorm(bmu[iter], sdbm.precision);
                 ###############
                  bmu[iter] <-  (1-delta[trt[iter]+1])*u[trt[iter]+1]+
delta[trt[iter]+1]*(u[1]+u[2])/2;
               }
               for(iter in 1:4){
                     bta[iter] ~ dnorm(0, 0.000001);
               }
               for (iter in 1:2){
                    delta[iter] ~ dbeta(0.0001,0.0001);
                    u[iter] ~ dnorm(0,000001);
               }
               sdbmd.precision ~ dgamma(0.00001,0.00001);
               sdbm.precision ~ dgamma(0.00001,0.00001);
            }
==================Data


list(biomarker= structure(.Data= c(-1.6332505600E-01, -9.9820335000E-02,
4.0692602900E-01, -8.2650000000E-03, -6.6739468400E-01, -1.9444291700E-01,
-2.5035248500E-01, -1.5926000000E-02, 2.5951119500E-01, 3.1943077100E-01,
2.1915456300E-01, -7.0600000000E-04, -1.1407237740E+00, -4.7155522000E-01,
-1.4495155440E+00, 1.7036000000E-02, -1.9105523700E-01, -5.6980210000E-03,
-8.4883829700E-01, 4.9390000000E-03), .Dim=c(5, 4)), bmd=c(-4.9440000000E-03,
3.5207000000E-02, 2.5975000000E-02, 2.7394000000E-02, 7.5010000000E-03),
trt=c(1.0000000000E+00, 0.0000000000E+00, 0.0000000000E+00, 1.0000000000E+00,
1.0000000000E+00), numSubj=5.0000000000E+00)

I think this is a question for some BUGS related mailing list rather 
than the R mailing lists.

Best,
Uwe Ligges



Haoda Fu wrote:
> Hi,
> 
> Is there anyone know if BUGS language allows the combination of variables
as response
> 
> such as
> Y[i] <- a*X1[i]+b*X2[i]
> Y[i] ~ dnorm(c,d)
> 
> 
> It seems doesn't work in my model. The problem is between two ######.
> The error message is 
>>  modelCheck("BayesBioMarker.BUGS")
> model is syntactically correct
>>  modelData(paste("BUGS_data.txt",sep=""))
> data loaded
>>  modelCompile(numChains=1)
> multiple definitions of node bm[1]
> 
> 
> ==================================>  model
>             {
>                for(iter in 1:numSubj){
>                  bmd[iter] ~ dnorm(u[trt[iter]+1],sdbmd.precision);
>                  ###############
>                  bm[iter] <-  inprod(biomarker[iter,1:4], bta[1:4])
>                  bm[iter] ~ dnorm(bmu[iter], sdbm.precision);
>                  ###############
>                   bmu[iter] <-  (1-delta[trt[iter]+1])*u[trt[iter]+1]+
delta[trt[iter]+1]*(u[1]+u[2])/2;
>                }
>                for(iter in 1:4){
>                      bta[iter] ~ dnorm(0, 0.000001);
>                }
>                for (iter in 1:2){
>                     delta[iter] ~ dbeta(0.0001,0.0001);
>                     u[iter] ~ dnorm(0,000001);
>                }
>                sdbmd.precision ~ dgamma(0.00001,0.00001);
>                sdbm.precision ~ dgamma(0.00001,0.00001);
>             }
> ==================> Data
> 
> 
> list(biomarker= structure(.Data= c(-1.6332505600E-01, -9.9820335000E-02,
4.0692602900E-01, -8.2650000000E-03, -6.6739468400E-01, -1.9444291700E-01,
-2.5035248500E-01, -1.5926000000E-02, 2.5951119500E-01, 3.1943077100E-01,
2.1915456300E-01, -7.0600000000E-04, -1.1407237740E+00, -4.7155522000E-01,
-1.4495155440E+00, 1.7036000000E-02, -1.9105523700E-01, -5.6980210000E-03,
-8.4883829700E-01, 4.9390000000E-03), .Dim=c(5, 4)), bmd=c(-4.9440000000E-03,
3.5207000000E-02, 2.5975000000E-02, 2.7394000000E-02, 7.5010000000E-03),
trt=c(1.0000000000E+00, 0.0000000000E+00, 0.0000000000E+00, 1.0000000000E+00,
1.0000000000E+00), numSubj=5.0000000000E+00)
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

> Is there anyone know if BUGS language allows the combination of 
> variables as response
> It seems doesn't work in my model. The problem is between two ######.
> >  modelCompile(numChains=1)
> multiple definitions of node bm[1]
>                  ###############
>                  bm[iter] <-  inprod(biomarker[iter,1:4], bta[1:4])
>                  bm[iter] ~ dnorm(bmu[iter], sdbm.precision);
>                  ###############
The problem seems to be that you've defined bm[iter] as a logical node 
(<-), then redefined it as a stochastic node (~).  If you need more help, 
try the BUGS mailing list.  (See 
http://www.mrc-bsu.cam.ac.uk/bugs/overview/list.shtml)

Regards,
Richie.

Mathematical Sciences Unit
HSL


------------------------------------------------------------------------
ATTENTION:

This message contains privileged and confidential inform...{{dropped:22}}