Why the stan code of two-way random effects wiener model is error

#1

The two-way random effects wiener model is
y_{ij} \sim N(\mu_{i}t_{ij}, \sigma_{i}^{2}t_{ij}),\mu_{i}\sim N(\eta,\tau^2),\sigma_{i}^{2}\sim invgamma(\alpha,\beta),i=1,\dots,N,j=1,\dots,M
where matrix t and t are observation data.

I have implement this model in OpenBUGs software with code

model
	{
		for( i in 1 : n) {
			for( j in 1 : m) {
				dy[i , j] ~ dnorm(mu1[i , j], sigma21[i,j])
				mu1[i , j] <- mu[i]*dt[i,j]
				sigma21[i,j] <- 1/(dt[i,j]*sigma22[i])
			}
			mu[i] ~ dnorm(eta, tau21)
			sigma2[i] ~ dgamma(alpha, beta)
			sigma22[i] <-  1/sigma2[i]
		}
		alpha ~ dgamma(0.001,0.001)
		beta ~ dgamma(0.0001,0.001)
		eta ~ dnorm(0.0,1.0E-6)
        tau21 <- 1/tau2		
		tau2 ~ dgamma(0.001,0.001)
	}

However, when I rewrite according to stan syntax with code

data {
int<lower=2> N;
int<lower=2> M;
matrix[M, N] y;
matrix[M, N] t;
}
parameters {
real eta;
real<lower=0> tau;
real<lower=1> alpa;
real<lower=0> beta;
real mu[N];
real<lower=0> sigma2[N];
}
model {
for (i in 1:N)
   mu ~ normal(eta,tau);
   sigma2 ~ inv_gamma(alpa, beta);
   for (j in 1:M)
     y[i,j] ~ normal(mu*t[i,j], sigma2^.5*t[i,j]);
 }

it seems not work. How can I modify the above stan code?

#2

mu[i] ... sigma2[i]

What is the error message?

#3

Thanks for ahartikainen, the error message,
SYNTAX ERROR, MESSAGE(S) FROM PARSER:

image
image.png751×251 7.25 KB

#4

Oh, you need to wrap multiline for -loops in {}

for (i in 1:N) {
    ....
    ....
}

Also, on RStan there was a problem with i as a variable name, but I think that is fixed.

#5

Also, this is unrelated, but given that your array is y[M,N] but you are indexing it y[i,j] with i in 1:N and j in 1:M, something is probably backward.