R jags.model节点中的错误与父节点不一致

我用一个β-二项n-混合模型来估计丰度。我编写的代码在模拟数据集和自然数据集上都运行良好。但是，当我运行一个特定的数据集时，这个错误会在jags.model（等等）中弹出error。节点N[1]中的error与父节点不一致我想知道为什么它与此数据集而不是其他数据集有问题。数据集中有相当数量的0不起作用，但数据集中有更多起作用的0也有相当数量的0。我已经使用了以前的值，范围从0.001到10。数据结构和模型如下所示

n.site <- 6 
R <- n.site #number of sites
T <- 10 #number of replicate counts
y <- array(dim = c(R, T)) #null array

y <- array(sur.2019$Num_total, dim = c(R, T)) #populate y
C <- c(y)
C <- as.numeric(C) 


site = 1:R              
site.p <- rep(site, T)

sink("Model.txt")
cat("
    model {
    
    # Priors
    lam~dgamma(0.01,0.01)
    alpha.p~dgamma(0.01,0.01)
    beta.p~dgamma(0.01,0.01)
    
    # Likelihood
    #the next four lines are used to model N as a zero-truncated distribution
    probs[R+1]<- 1-sum(probs[1:R])
    for(i in 1:R){
    probs[i]<- exp(-lam)*(pow(lam,x[i]))/(exp(logfact(x[i])) * (1-exp(-lam)))
    N[i] ~ dcat(probs[])}
    
    # Observation model for replicated counts
    for (i in 1:n) {                   
    C[i] ~ dbin(p[i], N[site.p[i]])
    p[i]~dbeta(alpha.p,beta.p)
    
    # Assess model fit using Chi-squared discrepancy
    # Compute fit statistic for observed data
    eval[i] <- p[i]*N[site.p[i]]
    E[i] <- pow((C[i] - eval[i]),2) / (eval[i] + 0.5)
    # Generate replicate data and compute fit stats
    C.new[i] ~ dbin(p[i], N[site.p[i]])
    E.new[i] <- pow((C.new[i] - eval[i]),2) / (eval[i] + 0.5)
    
    } # ends i loop
    
    
    # Derived and other quantities
    totalN <- sum(N[])  # Estimate abundance across all sites
    mean.abundance <- lam #mean expected abundance per plot
    p.derived<-alpha.p/(alpha.p+beta.p) #derived detection probability
    rho.derived<-1/(alpha.p+beta.p+1)  #correlation coefficient
    
    
    fit <- sum(E[])
    fit.new <- sum(E.new[])
    
    }
    ",fill = TRUE)
sink()

R = nrow(y)
T = ncol(y)
n = dim(y)[1] * dim(y)[2]#number of observations (sites*surveys)

nmm.data <- list(C = C, n=n, R = R, site.p = site.p, x=1:R)

# Initial values
Nst <- apply(y, 1, max) + 1 #changed from apply(y, 1, max) + 1
Nst[is.na(Nst)] <- 1
inits <- function(){list(N = Nst, lam = runif(1, 1, 7),alpha.p=runif(1,0.5,1), beta.p=runif(1,0.5,1))}

# Define parameters to be monitored
params <- c("totalN", "mean.abundance", "lam", "p.derived", "rho.derived", "fit", "fit.new","alpha.p","beta.p")

# MCMC settings
ni <- 14000
nt <- 1
nb <- 4000
nc <- 3



abun.1   <- jags(data       = nmm.data,
                 parameters = params,
                 inits      = inits,
                 model      = "model.txt",
                 n.thin     = nt, 
                 n.chains   = nc,
                 n.burnin   = nb,
                 n.iter     = ni)

abun.1.mcmc <- as.mcmc(abun.1)

嗯，数据就是这样

list(C = c(0, 0, 0, 0, 0, 2, 0, 1, 0, 0, 0, 0, 0, 3, 0, 1, 0, 
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 1, 1, 
0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 
0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0), n = 258L, R = 6L, site.p = c(1L, 
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 
6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 
4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 
6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 
4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 
6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 
4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 
6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 
4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 
6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 
4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 
6L), x = 1:6)

---

仔细检查您的代码后，我能够使用

nmm.data

训练贝叶斯模型。我发现了两个问题。您的优先级可能会产生非常小的值，这会在模型中产生问题。你可以改变，但我没有。第二个问题是你的初始值。它们太大了，这会影响链条的性能。由于先验值较小，初始值较大，因此使用大量迭代可能会搞乱一切。因此，当我使用这个麻烦的数据集时，我所做的是减少迭代次数（而且您的数据很小，当您有大数据作为其他列表时，问题不会出现）。这里是代码，在最后一部分，我将为您提供一些提取结果的选项

首先是模型（对此没有更改，但添加了固定的随机种子以保持结果相似）：

您将创建对象

m1.mcmc

，并在

res1

中保存每个参数链的结果。我添加了下一个输出的示例

d1

（因为输出很大，所以只有一些行

head（d1）

）：

---

因此，当您使用较大的数据时，链会收敛，而使用较小的数据时，如

nmm.data

请注意初始值和迭代次数。

请您

dput（nmm.data）

和

dput（hmm.data）

并将输出粘贴到您的问题中好吗？？根据您的分享，我们无法复制您的问题。还可以共享执行马尔可夫链检查所有规范的代码：）我已经添加了用于运行JAGS模型的其余代码，并在每个数据集上添加了来自dput的out。如果您需要其他信息，请告诉我，再次感谢您查看此代码。非常感谢！我不熟悉JAGS和贝叶斯统计，所以这对你组织和解释它的方式从来没有帮助。再次感谢您抽出时间来做这件事！

hmm.data <- list(C = C, n=n, R = R, site.p = site.p, x=1:R)
hmm.data

$C
  [1] 0 0 0 0 0 2 0 1 0 0 0 0 0 3 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 1 0 0 0 0 1 1 0 0 0 0 0 0 1 1 0 1 0 0 0 1 0 0
 [55] 0 0 0 1 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
[109] 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
[163] 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
[217] 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

$n
[1] 60

$R
[1] 6

$site.p
  [1] 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
 [55] 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
[109] 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
[163] 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
[217] 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6

$x
[1] 1 2 3 4 5 6

list(C = c(7, 0, 1, 0, 0, 0, 2, 0, 0, 0, 0, 1, 3, 2, 0, 0, 0, 
0, 1, 0, 0, 0, 0, 0, 5, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0), n = 60L, R = 6L, site.p = c(1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 
3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 
1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 
5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 
3L, 4L, 5L, 6L), x = 1:6)

list(C = c(0, 0, 0, 0, 0, 2, 0, 1, 0, 0, 0, 0, 0, 3, 0, 1, 0, 
0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 1, 1, 
0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 
0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 0), n = 258L, R = 6L, site.p = c(1L, 
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 
6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 
4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 
6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 
4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 
6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 
4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 
6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 
4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 
6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 
4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 
6L), x = 1:6)

library(rjags)
set.seed(123)
#Code for model
mymod <- "

model {
    
    # Priors
    lam~dgamma(0.01,0.01)
    alpha.p~dgamma(0.01,0.01)
    beta.p~dgamma(0.01,0.01)
    
    # Likelihood
    #the next four lines are used to model N as a zero-truncated distribution
    probs[R+1]<- 1-sum(probs[1:R])
    for(i in 1:R){
    probs[i]<- exp(-lam)*(pow(lam,x[i]))/(exp(logfact(x[i])) * (1-exp(-lam)))
    N[i] ~ dcat(probs[])}
    
    # Observation model for replicated counts
    for (i in 1:n) {                   
    C[i] ~ dbin(p[i], N[site.p[i]])
    p[i]~dbeta(alpha.p,beta.p)
    
    # Assess model fit using Chi-squared discrepancy
    # Compute fit statistic for observed data
    eval[i] <- p[i]*N[site.p[i]]
    E[i] <- pow((C[i] - eval[i]),2) / (eval[i] + 0.5)
    # Generate replicate data and compute fit stats
    C.new[i] ~ dbin(p[i], N[site.p[i]])
    E.new[i] <- pow((C.new[i] - eval[i]),2) / (eval[i] + 0.5)
    
    } # ends i loop
    
    
    # Derived and other quantities
    totalN <- sum(N[])  # Estimate abundance across all sites
    mean.abundance <- lam #mean expected abundance per plot
    p.derived<-alpha.p/(alpha.p+beta.p) #derived detection probability
    rho.derived<-1/(alpha.p+beta.p+1)  #correlation coefficient
    
    
    fit <- sum(E[])
    fit.new <- sum(E.new[])
    
    }
    "

#Data
#Values
n.site <- 6 
R <- n.site #number of sites
T <- 10 #number of replicate counts
y <- array(dim = c(R, T)) #null array
# Initial values
Nst <- apply(y, 1, max) + 1 #changed from apply(y, 1, max) + 1
Nst[is.na(Nst)] <- 1
inits <- function(){list(N = Nst, lam = runif(1, 1, 7),
                         alpha.p=runif(1,0.1,0.5),
                         beta.p=runif(1,0.5,1))}

# MCMC settings using nmm.data
ni <- 14000
nt <- 1
nb <- 4000
nc <- 3
#Model
m1 <- jags.model(file=textConnection(mymod),
           data=nmm.data,n.chains=3,inits=inits(),quiet = T)
#Update
update(m1, n.iter=1000,progress.bar = "none")

#Parameters
params <- c("totalN", "mean.abundance", "lam", "p.derived",
            "rho.derived", "fit", "fit.new","alpha.p","beta.p")
#Extract results
res1 <- coda.samples(m1,variable.names=params,n.iter=1000)
#Code for extracting
m1.mcmc <- as.mcmc(m1)
#Example
d1 <- as.data.frame(res1[[1]])

     alpha.p    beta.p      fit  fit.new      lam mean.abundance  p.derived rho.derived totalN
1 0.04932755 0.5086096 3.390977 3.765055 2.517308       2.517308 0.08841059   0.6418744     16
2 0.06047980 0.5160444 2.218640 3.986287 2.453345       2.453345 0.10490418   0.6343068     18
3 0.05784805 0.4795353 1.286302 2.862089 3.213626       3.213626 0.10764764   0.6504558     17
4 0.05672757 0.3837686 2.980358 2.260288 1.679854       1.679854 0.12878106   0.6942052     15
5 0.05653906 0.8254216 6.293583 2.671746 3.382379       3.382379 0.06410611   0.5313607     15
6 0.06015773 0.5851066 5.566315 2.555359 5.338035       5.338035 0.09322960   0.6078051     26