"expected right parenthesis error pos" using OpenBUGS

Question

I'm trying to run a simple linear regression analysis using "R2OpenBUGS" package in R. when i run the "bugs" command, i face with error.

After adding "debug = T" to the command line, i get this error in OpenBUGS:

expected right parenthesis error pos 130

here's my codes in R:

library(R2OpenBUGS)
library(coda)

MODEL <- function() {
  for (i in 1:N) {
    y[i] ~ dnorm(mu[i], s2)  
    mu[i] <- b0+b1*(x[i])      
  }
  b0 ~ dnorm(0, 1/9)           
  b1 ~ dnorm(0, 1/9)         
  s2 ~ dgamma(3, 1)            
  s2 <- 1/s2
}
write.model (model,"MODEL.txt")
INIT <- function() {
  list(b0 = runif(1,-9, 9), 
       b1 = runif(1,-9, 9),
       s2 = rexp(1,1/3))
}

DATA=list(y=c(15,33,26,21,39,40,14,38,20,32),
          x=c(3,7,6,3,9,9,2,9,5,6),N=10)      


BUGS=bugs(data = DATA, inits = INIT, 
          parameters.to.save = c("b0", "b1", "s2"), 
          model.file = "MODEL.txt", n.chains = 1,
          n.iter = 10000, n.burnin = 100, codaPkg=TRUE, debug = T)

Answer 1

You need to set the precision parameters of b0 and b1 to 0.11, not 1/9. The precision parameter must be a floating point number or OpenBUGS will return an error.

In addition, you have defined s2 as a random variable and then assigned a value to it. BUGS does not like that either. You should create a new variable tau to which to assign the value of the inverse of s2. Your model code should read as follows:

MODEL <- function() {
  for (i in 1:N) {
    y[i] ~ dnorm(mu[i], tau)  
    mu[i] <- b0+b1*(x[i])      
  }
  b0 ~ dnorm(0, 0.11)           
  b1 ~ dnorm(0, 0.11)         
  s2 ~ dgamma(3, 1)            
  tau <- 1/s2
}