在尝试调整工作WinBUGS模型并使用R2WinBUGS将其缓解为R时，我收到了几条错误消息。
我认为这与inits的规范有关，但无法解决问题。
第一个是使用inits1的错误消息：
list(A=1, d=c(NA,0,0,0,0), mu=c(0,0,0,0))
Error in bugs(mydata, inits = inits1, model.file = "mtcfe.txt", parameters = c("or"), :
Number of initialized chains (length(inits)) != n.chains
在阅读Uwe Liggers "List containing lists solution"建议的修正后，我将inits1修改为inits2：
inits2 <- list(list(A=1, d=c(NA,0,0,0,0), mu=c(0,0,0,0)))
收到错误：
Error in bugs.run(n.burnin, bugs.directory, WINE = WINE, useWINE = useWINE, : Look at the log file and try again with 'debug=TRUE' to figure out what went wrong within Bugs.
我也尝试过AndyC在这篇文章"getting-winbugs-leuk-example-to-work-from-r-using-r2winbugs"上提出的修正案。将inits更改为：
inits4 <- function(){list(list(A=1, d=c(NA,0,0,0,0), mu=c(0,0,0,0)))}
收到错误：
Error in bugs.run(n.burnin, bugs.directory, WINE = WINE, useWINE = useWINE, : Look at the log file and try again with 'debug=TRUE' to figure out what went wrong within Bugs.
这是我对R2WinBUGS的尝试。
注意，我在代码中包含了几个不起作用的inits：
work.dir <- "removed from example"
setwd(work.dir)
getwd() # check working directory
# Load Package
library(R2WinBUGS)
# Read data
mydata <- list(nt=5,
ns=4,
r=structure(
.Data = c(2506,7834,6729,2139,
2548,7860,6710,4418),
.Dim = c(4,2)),
n=structure(
.Data = c(2697, 8212, 7266, 2333,
2701,8280,7257,4687),
.Dim= c(4,2)),
t=structure(
.Data = c( 1,1,1,1,
2,3,4,5),
.Dim = c(4,2)),
na=structure(
.Data = c(2,2,2,2))
)
bugs.data(mydata)
# Set initial values
inits <- function(){list(A=1, d=c(NA,0,0,0,0), mu=c(0,0,0,0))}
#inits2 <- list(A=1, d=c(NA,0,0,0,0), mu=c(0,0,0,0))
#inits3 <- function(){list(inits2,inits2)}
#inits4 <- function(){list(inits,inits)}
#inits5 <- list(inits,inits)
# CALL WinBUGS AND SAVE RESULTS IN VARIABLE out.re
out.re <- bugs(mydata,inits=inits1, # load data and initial values
model.file="mtcfe.txt", # file with model to run
parameters=c("or"),
n.thin=1,
n.chains=1, n.iter=1500, n.burnin=500,
bugs.directory=bd,
working.directory=work.dir,
debug=TRUE)
print(out.re,digits=4) #lists all results as in WinBUGS stats(*)
这是一个有效的WinBUGS代码，改编自布里斯托尔和莱斯特大学举办的WinBUGS课程：
# Binomial likelihood, logit link, MTC
# Fixed effect model
model{ # *** PROGRAM STARTS
for(i in 1:ns){ # LOOP THROUGH STUDIES
mu[i] ~ dnorm(0,.0001) # vague priors for all trial baselines
for (k in 1:na[i]) { # LOOP THROUGH ARMS
r[i,k] ~ dbin(p[i,k],n[i,k]) # binomial likelihood
logit(p[i,k]) <- mu[i] + d[t[i,k]]-d[t[i,1]] # model for linear predictor
rhat[i,k] <- p[i,k] * n[i,k] # expected value of the numerators
dev[i,k] <- 2 * (r[i,k] * (log(r[i,k])-log(rhat[i,k])) #Deviance contribution
+ (n[i,k]-r[i,k]) * (log(n[i,k]-r[i,k]) - log(n[i,k]-rhat[i,k])))
}
resdev[i] <- sum(dev[i,1:na[i]]) # summed residual deviance contribution for this trial
}
totresdev <- sum(resdev[]) #Total Residual Deviance
d[1]<- 0 # treatment effect is zero for reference treatment
for (k in 2:nt) { d[k] ~ dnorm(0,.0001) } # vague priors for treatment effects
# pairwise ORs and LORs for all possible pair-wise comparisons
for (c in 1:(nt-1)) { for (k in (c+1):nt) {
or[c,k] <- exp(d[k] - d[c])
lor[c,k] <- (d[k]-d[c])
}
}
# ranking
for (k in 1:nt) {
rk[k] <- nt+1-rank(d[],k) # assumes events are “good”
# rk[k] <- rank(d[],k) # assumes events are “bad”
best[k] <- equals(rk[k],1) #calculate probability that treat k is best
}
# Absolute effects
A ~ dnorm(-2.6,precA)
precA <- pow(0.38,-2) # prior precision for Treatment A, sd=0.38 on logit scale
for (k in 1:nt) { logit(T[k]) <- A + d[k] }
} # *** PROGRAM ENDS
#Inits
list(A=1, d=c(NA,0,0,0,0), mu=c(0,0,0,0))
#Data
list(nt=5.00000E+00, ns=4.00000E+00, r= structure(.Data= c(2.50600E+03, 2.54800E+03, 7.83400E+03, 7.86000E+03, 6.72900E+03, 6.71000E+03, 2.13900E+03, 4.41800E+03), .Dim=c(4, 2)), n= structure(.Data= c(2.69700E+03, 2.70100E+03, 8.21200E+03, 8.28000E+03, 7.26600E+03, 7.25700E+03, 2.33300E+03, 4.68700E+03), .Dim=c(4, 2)), t= structure(.Data= c(1.00000E+00, 2.00000E+00, 1.00000E+00, 3.00000E+00, 1.00000E+00, 4.00000E+00, 1.00000E+00, 5.00000E+00), .Dim=c(4, 2)), na=c(2.00000E+00, 2.00000E+00, 2.00000E+00, 2.00000E+00))