I am having trouble initializing a bayesian model, with multivariate normal
likelihood.
*model{*
for (i in 1:N) {
mu1[i]<-(pow(10,(pka1-ph[i]))*da1[2]+da1[1])/(1+pow(10,(pka1-ph[i])))+K[1]
}
for (i in 1:N) {
mu2[i]<-(pow(10,(pka2-ph[i]))*da2[2]+da2[1])/(1+pow(10,(pka2-ph[i])))+K[2]
}
Y[1:N,1]<-y1
Y[1:N,2]<-y2
MU[1,1:N]<-mu1
MU[2,1:N]<-mu2
Y ~ dmnorm(MU,SIGMA)
SIGMA ~ dwish(R,2)
K[1] ~ dnorm(0,t2[1])
t2[1]<-1/p[1]
p[1] ~ dgamma(2,0.5)
K[2] ~ dnorm(0,t2[2])
t2[2]<-1/p[2]
p[2] ~ dgamma(2,0.5)
pka1 ~ dnorm(A,1)
da1[1] ~ dunif(0,10)
da1[2] ~ dunif(0,10)
pka2 ~ dnorm(A,1)
da2[1] ~ dunif(0,10)
da2[2] ~ dunif(0,10)
A ~ dunif(0,14)
*}*
y1 and y2 are vectors of length N, whilst the matrix R is the N-dim identity
matrix.
When attempting to initialize, I get the following error:
/ RUNTIME ERROR:
Non-conforming parameters in distribution dmnorm/
--
View this message in context:
http://r.789695.n4.nabble.com/rjags-dmnorm-error-tp4620137.html
Sent from the R help mailing list archive at Nabble.com.
______________________________________________
R-help@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.
