model {
for (n in 1:3) { #定义有三个先验的循环
for (j in 1:K) {
for (i in 1:M) {
alpha[i,j]~dnorm(mu[n,j], tau[n])
}
mu[n,j] <- t[n,j]+a[n,j]
t[n,j]~dnorm(m[n],V[n])
a[n,j]~dnorm(0,sigma[n])
}
}
# 先验分布1
tau[1]~dgamma(100,0.0075)
sigma[1]~dgamma(10,0.00033)
m[1]~dnorm(0,400)
V[1]~dgamma(100,0.0075)
谢谢回复,但是在winbugs的manual中在Tricks: Advanced use of the bugs language这一章就有提到assessing sensitivity to prior assumptions,其中还有一个例子,就是用一个循环来做,我是参考他来写的,就是对结果和单独做相差很多。