Question

我想构建一个像

这样的模型

Model3 <- function() {
  # Likelihood of the model  
  for (i in 1:n) {
    response2[i] ~ dbern(p[i])
    logit(p[i]) <- m[i]
    m[i] <- mfe[i]
    mfe[i] <- beta[1] + f[i]  

    # the f functions or the if-else step like
    f_branch[i, 2]  <- alpha 
    f_branch[i, 1]  <- 1+alpha*exp(-((1/phi)*X2[i,1])^2) 

    # the decision
    if_branch[i] <- 1 + step(-(X2[i,1] - 2)) 
    f[i] <- f_branch[i, if_branch[i]]
    f[i] <- f_branch[i, if_branch[i]]
  }

  # Prediction step
  for(ii in 1:ndist) {  
    if_branch2[ii] <- 1 + step(-(Dist[ii] - 2))     
    fhat[ii] <- f_branch2[ii, if_branch2[ii]]    
    fhat[ii] <- f_branch2[ii, if_branch2[ii]]     
    f_branch2[ii, 2] <- alpha 
    f_branch2[ii, 1] <- 1 + alpha*exp(-((1/phi)*Dist[ii])^2)
  }

  # Prior distribution of the fixed effects parameters
  beta[1] ~ dnorm(0, 1.0E-6)}

  # priors for the f function
  alpha ~ dgamma(1, 1)
  phi ~ dunif(0, 2)
}

我收到错误：Attempt to redefine node f[1]

Answer 1

在这种情况下，错误消息非常明显 - 您定义f[1]，然后在下一行再次定义它，因为您重复了这一行：

f[i] <- f_branch[i, if_branch[i]]

删除该行的第二个实例，该错误应该消失。

您还复制了该行：

fhat[ii] <- f_branch2[ii, if_branch2[ii]]

所以你也需要删除它的第二个实例。

WinBugs if else使用step函数

1 个答案: