------------------------------------------------------------------------
JAGS的錯誤訊息
[軟體程式類別]:
JAGS
[程式問題]:
Multilevel analysis
[軟體熟悉度]:
中
[問題敘述]:
各位大哥大姐：
小弟我跑以下的多層次模型，JAGS一直給我如下錯誤訊息：
model{
for (i in 1:N){
y.2010[i] ~ dbin(pi[i], 1)
pi[i] <- exp(Xbeta[i]+epsilon[i])/(1+exp(Xbeta[i]+epsilon[i]))
Xbeta[i] <- b.Con + b.pid_inc*pidinc.ok[i] +
b.soc*soc.ok[i] + a.age[age.ok[i]] + a.edu[edu.ok[i]] +a.county[county.ok[i]]
epsilon[i] ~ dnorm(e.gamma, e.tau)
}
e.gamma <- 0
b.Con ~ dnorm(2, 0.0001)
b.pid_inc ~ dnorm(2, 0.0001)
b.soc ~ dnorm(2, 0.0001)
for (j in 1:n.age) {a.age[j] ~ dnorm(0, tau.age)}
for (j in 1:n.edu) {a.edu[j] ~ dnorm(0, tau.edu)}
for (j in 1:n.county) {
a.county[j] ~ dnorm(a.county.hat[j], tau.county)
a.county.hat[j] <- a.area[area3[j]]+beta.soc*soc.i[j]
}
for (j in 1:n.area) {a.area[j] ~ dnorm(0, tau.area)}
beta.soc ~ dnorm(0, .0001)
tau.age <- pow(sigma.age, -2)
tau.edu <- pow(sigma.edu, -2)
tau.county <- pow(sigma.county, -2)
tau.area <- pow(sigma.area, -2)
sigma.age ~ dunif (0, 100)
sigma.edu ~ dunif (0, 100)
sigma.county ~ dunif (0, 100)
sigma.area ~ dunif (0, 100)
e.tau ~ dgamma(0.01, 0.01)
}
Error : 外部函數呼叫時不能有 NA/NaN/Inf (引數 1)
此外: Warning messages:
1: glm.fit: algorithm did not converge
2: step size truncated due to divergence
請問一下如何解決？並且讓方程式converge? Any help is highly appreciated.
--
※ 發信站: 批踢踢實業坊(ptt.cc)
◆ From: 140.119.244.172