WinBUGS问题求解

ctfy · 2013年1月24日

模型如下：

<br />
model<br />
{<br />
    for (i in 1:n) {<br />
        y[i] ~ dbern(p[i])<br />
        logit(p[i]) <- b[1] + b[2] * x[i, 1] + b[3] * x[i, 2]<br />
        x[i, 1:2] ~ dmnorm(m[], t.m[, ])<br />
    }<br />
    for (i in 1:3) {<br />
        b[i] ~ dnorm(0.00000E+00, 1.00000E-04)<br />
    }<br />
    m[1] <- 0.00000E+00<br />
    m[2] <- 0.00000E+00<br />
    t.m[1, 1] <- 1.00000E-04<br />
    t.m[1, 2] <- 0.00000E+00<br />
    t.m[2, 1] <- 0.00000E+00<br />
    t.m[2, 2] <- 1.00000E-04<br />
}<br />

数据如下：

[data]

list(n=5.00000E+01, y=c(0.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00), x= structure(.Data= c(-5.60476E-01, NA, -2.30177E-01, -2.85468E-02, NA, -4.28705E-02, 7.05084E-02, 1.36860E+00, 1.29288E-01, -2.25771E-01, 1.71506E+00, 1.51647E+00, NA, -1.54875E+00, NA, 5.84614E-01, NA, 1.23854E-01, -4.45662E-01, NA, NA, 3.79639E-01, 3.59814E-01, -5.02323E-01, 4.00771E-01, NA, 1.10683E-01, NA, -5.55841E-01, -1.07179E+00, 1.78691E+00, 3.03529E-01, 4.97850E-01, NA, NA, NA, 7.01356E-01, 9.22267E-01, -4.72791E-01, 2.05008E+00, -1.06782E+00, NA, NA, -2.30917E+00, NA, 1.00574E+00, NA, NA, -6.25039E-01, -6.88009E-01, -1.68669E+00, 1.02557E+00, 8.37787E-01, NA, NA, -1.22072E+00, -1.13814E+00, 1.81303E-01, 1.25381E+00, NA, 4.26464E-01, NA, -2.95071E-01, 3.85280E-01, 8.95126E-01, -3.70660E-01, 8.78133E-01, 6.44377E-01, 8.21581E-01, -2.20487E-01, 6.88640E-01, 3.31782E-01, 5.53918E-01, NA, -6.19117E-02, 4.35181E-01, -3.05963E-01, -3.25932E-01, -3.80471E-01, 1.14881E+00, NA, 9.93504E-01, NA, 5.48397E-01, NA, 2.38732E-01, NA, NA, 1.20796E+00, 1.36065E+00, -1.12311E+00, -6.00260E-01, NA, 2.18733E+00, NA, 1.53261E+00, 7.79965E-01, NA, NA, -1.02642E+00), .Dim=c(50, 2)))

[/data]

问题如下：

编译时出现：

model is syntactically correct

data loaded

****** Sorry something went wrong in procedure Updater.AllocateSpecial in module UpdaterGLM ******

multiple definitions of node p[1]

怎么也没有看出为啥p这个节点多次定义了

网上找了好久也没有看到相应的解答

而且如果将上面的GLM改成普通线性模型，则没有任何错误。

不知道什么原因？求大牛解答。谢谢~

ctfy · 2013年1月25日

另外，同样的方式，对于1个协变量的情况，则一点儿问题都没有，例如

<br />
    model<br />
    {<br />
        for (i in 1:n) {<br />
            y[i] ~ dbern(p[i])<br />
            logit(p[i]) <- b0 + b1 * x[i]<br />
            x[i] ~ dnorm(mu, tau.m)<br />
        }<br />
        mu ~ dnorm(0.00000E+00, 1.00000E-04)<br />
        tau.m ~ dgamma(1.00000E-04, 1.00000E-04)<br />
        b0 ~ dnorm(0.00000E+00, 1.00000E-04)<br />
        b1 ~ dnorm(0.00000E+00, 1.00000E-04)<br />
    }<br />

数据如下：

[data]

list(n=5.00000E+01, y=c(1.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00), x=c(-5.60476E-01, -2.30177E-01, 1.55871E+00, 7.05084E-02, 1.29288E-01, 1.71506E+00, 4.60916E-01, -1.26506E+00, -6.86853E-01, -4.45662E-01, 1.22408E+00, 3.59814E-01, 4.00771E-01, 1.10683E-01, NA, 1.78691E+00, 4.97850E-01, -1.96662E+00, 7.01356E-01, -4.72791E-01, NA, NA, -1.02600E+00, -7.28891E-01, -6.25039E-01, -1.68669E+00, NA, 1.53373E-01, NA, 1.25381E+00, NA, -2.95071E-01, 8.95126E-01, 8.78133E-01, 8.21581E-01, 6.88640E-01, NA, NA, -3.05963E-01, -3.80471E-01, NA, -2.07917E-01, NA, 2.16896E+00, 1.20796E+00, -1.12311E+00, -4.02885E-01, -4.66655E-01, 7.79965E-01, NA))

[/data]

初始值如下：

[data]

list(mu=0.00000E+00, tau.m=1.00000E-01, b0=0.00000E+00, b1=1.00000E+00, x=c( NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, 0.00000E+00, NA, NA, NA, NA, NA, 0.00000E+00, 0.00000E+00, NA, NA, NA, NA, 0.00000E+00, NA, 0.00000E+00, NA, 0.00000E+00, NA, NA, NA, NA, NA, 0.00000E+00, 0.00000E+00, NA, NA, 0.00000E+00, NA, 0.00000E+00, NA, NA, NA, NA, NA, NA, 0.00000E+00))

[/data]

ctfy · 2013年1月25日

以下情况也可以运行：

模型如下：

<br />
    model<br />
    {<br />
        for (i in 1:n) {<br />
            y[i] ~ dbern(p[i])<br />
            logit(p[i]) <- b[1] + b[2] * x[i, 1] + b[3] * x[i, 2]<br />
            x[i, 1] ~ dnorm(mu, tau.m)<br />
            x[i, 2] ~ dnorm(mu, tau.m)<br />
        }<br />
        for (i in 1:3) {<br />
            b[i] ~ dnorm(0.00000E+00, 1.00000E-04)<br />
        }<br />
        mu ~ dnorm(0.00000E+00, 1.00000E-04)<br />
        tau.m ~ dgamma(1.00000E-04, 1.00000E-04)<br />
    }<br />

数据如下：

[data]

list(n=5.00000E+01, y=c(0.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00), x= structure(.Data= c(-5.60476E-01, NA, -2.30177E-01, -2.85468E-02, NA, -4.28705E-02, 7.05084E-02, 1.36860E+00, 1.29288E-01, -2.25771E-01, 1.71506E+00, 1.51647E+00, NA, -1.54875E+00, NA, 5.84614E-01, NA, 1.23854E-01, -4.45662E-01, NA, NA, 3.79639E-01, 3.59814E-01, -5.02323E-01, 4.00771E-01, NA, 1.10683E-01, NA, -5.55841E-01, -1.07179E+00, 1.78691E+00, 3.03529E-01, 4.97850E-01, NA, NA, NA, 7.01356E-01, 9.22267E-01, -4.72791E-01, 2.05008E+00, -1.06782E+00, NA, NA, -2.30917E+00, NA, 1.00574E+00, NA, NA, -6.25039E-01, -6.88009E-01, -1.68669E+00, 1.02557E+00, 8.37787E-01, NA, NA, -1.22072E+00, -1.13814E+00, 1.81303E-01, 1.25381E+00, NA, 4.26464E-01, NA, -2.95071E-01, 3.85280E-01, 8.95126E-01, -3.70660E-01, 8.78133E-01, 6.44377E-01, 8.21581E-01, -2.20487E-01, 6.88640E-01, 3.31782E-01, 5.53918E-01, NA, -6.19117E-02, 4.35181E-01, -3.05963E-01, -3.25932E-01, -3.80471E-01, 1.14881E+00, NA, 9.93504E-01, NA, 5.48397E-01, NA, 2.38732E-01, NA, NA, 1.20796E+00, 1.36065E+00, -1.12311E+00, -6.00260E-01, NA, 2.18733E+00, NA, 1.53261E+00, 7.79965E-01, NA, NA, -1.02642E+00), .Dim=c(50, 2)))

[/data]

初值如下：

[data]

list(mu=0.00000E+00, tau.m=1.00000E-01, b=c(0.00000E+00, 1.00000E+00, 1.00000E+00), x= structure(.Data= c( NA, 0.00000E+00, NA, NA, 0.00000E+00, NA, NA, NA, NA, NA, NA, NA, 0.00000E+00, NA, 0.00000E+00, NA, 0.00000E+00, NA, NA, 0.00000E+00, 0.00000E+00, NA, NA, NA, NA, 0.00000E+00, NA, 0.00000E+00, NA, NA, NA, NA, NA, 0.00000E+00, 0.00000E+00, 0.00000E+00, NA, NA, NA, NA, NA, 0.00000E+00, 0.00000E+00, NA, 0.00000E+00, NA, 0.00000E+00, 0.00000E+00, NA, NA, NA, NA, NA, 0.00000E+00, 0.00000E+00, NA, NA, NA, NA, 0.00000E+00, NA, 0.00000E+00, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, 0.00000E+00, NA, NA, NA, NA, NA, NA, 0.00000E+00, NA, 0.00000E+00, NA, 0.00000E+00, NA, 0.00000E+00, 0.00000E+00, NA, NA, NA, NA, 0.00000E+00, NA, 0.00000E+00, NA, NA, 0.00000E+00, 0.00000E+00, NA), .Dim=c(50, 2)))

[/data]

这个时候协变量x1和x2是独立的先验。

============================

对于连续的响应变量，也没有问题。

模型如下：

<br />
    model<br />
    {<br />
        for (i in 1:n) {<br />
            y[i] ~ dlnorm(u[i], tau.u)<br />
            u[i] <- b[1] + b[2] * x[i, 1] + b[3] * x[i, 2]<br />
            x[i, 1:2] ~ dmnorm(m[], tau.m[, ])<br />
        }<br />
        for (i in 1:3) {<br />
            b[i] ~ dnorm(0.00000E+00, 1.00000E-04)<br />
        }<br />
        tau.u ~ dgamma(0.01, 0.01)<br />
        m[1] <- 0.00000E+00<br />
        m[2] <- 0.00000E+00<br />
        tau.m[1:2, 1:2] ~ dwish(t.m[, ], 2)<br />
        t.m[1, 1] <- 1.00000E-04<br />
        t.m[1, 2] <- 0.00000E+00<br />
        t.m[2, 1] <- 0.00000E+00<br />
        t.m[2, 2] <- 1.00000E-04<br />
    }<br />

数据如下：

[data]

list(n=5.00000E+01, y=c(8.65414E-01, 1.12487E+00, 5.24423E+00, 2.81524E+00, 1.17452E+00, 1.72255E+01, 8.21693E-01, 5.21954E-01, 8.12168E-01, 1.12307E+00, 5.43900E+00, 1.29219E+00, 1.55146E+00, 8.78418E-01, 3.83364E-01, 7.27281E+00, 2.61510E+00, 2.17445E-01, 4.72890E+00, 2.88193E+00, 3.49869E-01, 2.72516E-01, 1.04425E+00, 4.36516E-01, 4.26072E-01, 4.36151E-01, 2.59282E+00, 6.89375E-01, 5.29516E-01, 4.39433E+00, 2.26427E+00, 1.22845E+00, 2.31174E+00, 4.35802E+00, 1.91039E+00, 2.72710E+00, 4.45600E+00, 1.54100E+00, 8.07743E-01, 1.57474E+00, 1.09645E+00, 1.40779E+00, 4.02234E-01, 7.33391E+00, 8.09886E+00, 3.37779E-01, 3.51361E+00, 1.79529E+00, 2.53790E+00, 4.99779E-01), x= structure(.Data= c( NA, 2.53319E-01, -2.30177E-01, -2.85468E-02, 1.55871E+00, NA, 7.05084E-02, 1.36860E+00, 1.29288E-01, -2.25771E-01, 1.71506E+00, 1.51647E+00, 4.60916E-01, NA, -1.26506E+00, 5.84614E-01, -6.86853E-01, 1.23854E-01, NA, NA, 1.22408E+00, 3.79639E-01, 3.59814E-01, -5.02323E-01, NA, NA, NA, -1.01858E+00, -5.55841E-01, -1.07179E+00, 1.78691E+00, 3.03529E-01, NA, 4.48210E-01, NA, 5.30042E-02, 7.01356E-01, 9.22267E-01, -4.72791E-01, NA, NA, -4.91031E-01, -2.17975E-01, -2.30917E+00, -1.02600E+00, NA, NA, NA, -6.25039E-01, -6.88009E-01, -1.68669E+00, 1.02557E+00, NA, NA, 1.53373E-01, -1.22072E+00, -1.13814E+00, NA, NA, -1.38891E-01, NA, NA, -2.95071E-01, 3.85280E-01, 8.95126E-01, NA, 8.78133E-01, 6.44377E-01, 8.21581E-01, -2.20487E-01, 6.88640E-01, 3.31782E-01, NA, 1.09684E+00, -6.19117E-02, NA, -3.05963E-01, -3.25932E-01, -3.80471E-01, NA, -6.94707E-01, NA, -2.07917E-01, 5.48397E-01, -1.26540E+00, 2.38732E-01, NA, -6.27906E-01, 1.20796E+00, 1.36065E+00, -1.12311E+00, -6.00260E-01, -4.02885E-01, NA, -4.66655E-01, NA, NA, -2.35700E-01, -8.33691E-02, NA), .Dim=c(50, 2)))

[/data]

初值如下：

[data]

list(b=c(0.00000E+00, 1.00000E+00, 1.00000E+00), tau.m= structure(.Data= c(1.00000E-01, 0.00000E+00, 0.00000E+00, 1.00000E-01), .Dim=c(2, 2)), tau.u=1.00000E-01, x= structure(.Data= c(0.00000E+00, NA, NA, NA, NA, 0.00000E+00, NA, NA, NA, NA, NA, NA, NA, 0.00000E+00, NA, NA, NA, NA, 0.00000E+00, 0.00000E+00, NA, NA, NA, NA, 0.00000E+00, 0.00000E+00, 0.00000E+00, NA, NA, NA, NA, NA, 0.00000E+00, NA, 0.00000E+00, NA, NA, NA, NA, 0.00000E+00, 0.00000E+00, NA, NA, NA, NA, 0.00000E+00, 0.00000E+00, 0.00000E+00, NA, NA, NA, NA, 0.00000E+00, 0.00000E+00, NA, NA, NA, 0.00000E+00, 0.00000E+00, NA, 0.00000E+00, 0.00000E+00, NA, NA, NA, 0.00000E+00, NA, NA, NA, NA, NA, NA, 0.00000E+00, NA, NA, 0.00000E+00, NA, NA, NA, 0.00000E+00, NA, 0.00000E+00, NA, NA, NA, NA, 0.00000E+00, NA, NA, NA, NA, NA, NA, 0.00000E+00, NA, 0.00000E+00, 0.00000E+00, NA, NA, 0.00000E+00), .Dim=c(50, 2)))

[/data]

不知是什么原因？

YSU · 2013年1月26日

回复第1楼的 catfish：

一楼的模型我compile了，没问题。但是我不懂为什么把的x的分布参数: dmnorm(m[], t.m[, ]) 要确定死？我觉的起码var-cov部分要让模型来估计。

<br />
model<br />
{<br />
    for (i in 1:n) {<br />
        y[i] ~ dbern(p[i])<br />
        logit(p[i]) <- b[1] + b[2] * x[i, 1] + b[3] * x[i, 2]<br />
#        x[i, 1:2] ~ dmnorm(m[], t.m[, ])<br />
          x[i, 1:2] ~ dmnorm(m[], tau.m[, ])<br />
    }<br />
## Prior<br />
    for (i in 1:3) {<br />
        b[i] ~ dnorm(0.00000E+00, 1.00000E-04)<br />
    }<br />
    m[1] <- 0.00000E+00<br />
    m[2] <- 0.00000E+00<br />
#   t.m[1, 1] <- 1.00000E-04<br />
#    t.m[1, 2] <- 0.00000E+00<br />
#    t.m[2, 1] <- 0.00000E+00<br />
#    t.m[2, 2] <- 1.00000E-04<br />
        tau.m[1:2, 1:2] ~ dwish(t.m[, ], 2)<br />
        t.m[1, 1] <- 1.00000E-04<br />
        t.m[1, 2] <- 0.00000E+00<br />
        t.m[2, 1] <- 0.00000E+00<br />
        t.m[2, 2] <- 1.00000E-04<br />
}</p>
<p>############################<br />
## Data<br />
list(n=5.00000E+01, y=c(0.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 1.00000E+00, 0.00000E+00), x= structure(.Data= c(-5.60476E-01,     NA, -2.30177E-01, -2.85468E-02,     NA, -4.28705E-02, 7.05084E-02, 1.36860E+00, 1.29288E-01, -2.25771E-01, 1.71506E+00, 1.51647E+00,     NA, -1.54875E+00,     NA, 5.84614E-01,     NA, 1.23854E-01, -4.45662E-01,     NA,     NA, 3.79639E-01, 3.59814E-01, -5.02323E-01, 4.00771E-01,     NA, 1.10683E-01,     NA, -5.55841E-01, -1.07179E+00, 1.78691E+00, 3.03529E-01, 4.97850E-01,     NA,     NA,     NA, 7.01356E-01, 9.22267E-01, -4.72791E-01, 2.05008E+00, -1.06782E+00,     NA,     NA, -2.30917E+00,     NA, 1.00574E+00,     NA,     NA, -6.25039E-01, -6.88009E-01, -1.68669E+00, 1.02557E+00, 8.37787E-01,     NA,     NA, -1.22072E+00, -1.13814E+00, 1.81303E-01, 1.25381E+00,     NA, 4.26464E-01,     NA, -2.95071E-01, 3.85280E-01, 8.95126E-01, -3.70660E-01, 8.78133E-01, 6.44377E-01, 8.21581E-01, -2.20487E-01, 6.88640E-01, 3.31782E-01, 5.53918E-01,     NA, -6.19117E-02, 4.35181E-01, -3.05963E-01, -3.25932E-01, -3.80471E-01, 1.14881E+00,     NA, 9.93504E-01,     NA, 5.48397E-01,     NA, 2.38732E-01,     NA,     NA, 1.20796E+00, 1.36065E+00, -1.12311E+00, -6.00260E-01,     NA, 2.18733E+00,     NA, 1.53261E+00, 7.79965E-01,     NA,     NA, -1.02642E+00), .Dim=c(50, 2)))</p>
<p>##############################<br />
## initial estimates (one chain)<br />
list(b=c(0.1, 0.1, 0.1),  tau.m= structure(.Data= c(1.00000E+01, 0.00000E+00, 0.00000E+00, 1.00000E+01), .Dim=c(2, 2))<br />

ctfy · 2013年1月26日

回复第4楼的 YSU：谢谢你的答复，我只是想测试一下代码有没有问题，所以简单处理了一下。原来的代码跟你现在发的差不多。不过，我用你的代码编译还是不成功。我用的是OpenBUGS 3.2.2 rev1063 (2012-7-15)版本的，难道是有bug？

YSU · 2013年1月26日

回复第5楼的 catfish：我把四楼的模型走了一遍，运行正常呀。我用winbugs1.4. 没有用过Openbugs，无法重現你的错误信息

enita · 2013年1月27日

回复第3楼的 catfish：您好，您这个分析变量也是二分类的吗？如果是那么那个y在用R读取时是怎么编的语言，我现在也在做这个，可是在data loading时出现错误，可能是数据格式有问题，所以请教您，怎样在R中读取的数据。谢谢

ctfy · 2013年2月6日

回复第6楼的 YSU：恩，在WinBUGS1.4中运行没有错误，但在OpenBUGS中确实存在存在问题。我已经联系过OpenBUGS的作者，他对源代码作了一点小改动，估计下一个版本就不会有问题了。

ctfy · 2013年2月6日

回复第7楼的 enita：我没有太明白你问题的意思。就我给出的这个代码而言，主要是针对做二分类变量Logistic回归时会存在协变量部分缺失，然后假定协变量都是连续的且服从多元正态分布，这样可以通过Bayesian方法来估计含有缺失数值的模型参数值以及缺失数值的期望与方差。问题就在于目前版本的OpenBUGS不能很好的解决含有缺失值的多元正态协变量的参数估值。这点在WinBUGS1.4中没有问题。如果你单纯地想知道怎么用R来调用BUGS代码的话，可以读一下R2WinBUGS或者R2OpenBUGS中bugs函数的帮助问题，特别是example中的例子。

enita · 2013年3月1日

回复第9楼的 catfish：您好，我再Winbugs1.4中进行compile时，出现了invalid range specified错误，是什么原因呢？谢谢

enita · 2013年4月5日

回复第9楼的 catfish：回复第9楼的 catfish：您好，我在compile时也出现了multiple definitions of node pp的问题，不知道该怎么解决，您的问题是怎么解决的？希望您能帮帮我，谢谢