fixed myers_jags run, loads knitr_defaults

cboettig · May 29, 2013 · 783ed11 · 783ed11
1 parent 5e3f3f5
commit 783ed11
Show file tree

Hide file tree

Showing 2 changed files with 1,369 additions and 27 deletions.
diff --git a/inst/examples/BUGS/parametric-vs-nonparametric.Rmd b/inst/examples/BUGS/parametric-vs-nonparametric.Rmd
@@ -3,18 +3,10 @@
 Plotting and knitr options, (can generally be ignored)
 
 ```{r plotting-options, cache=FALSE, message=FALSE, warning=FALSE, include=FALSE}
-setwd("~/Documents/code/nonparametric-bayes/inst/examples/BUGS/")
-#library(knitcitations)
-library(ggplot2) 
-library(nonparametricbayes) 
-opts_chunk$set(tidy=FALSE, warning=FALSE, message=FALSE, cache=TRUE, comment=NA, fig.path="figure/parametric-"
-               fig.width=6, fig.height=4, cache.path="parametric-vs-nonparametric/", verbose=TRUE)
+source("~/.knitr_defaults.R")
 #opts_knit$set(upload.fun = socialR::flickr.url)
-theme_set(theme_bw(base_size=12))
-theme_update(panel.background = element_rect(fill = "transparent", colour = NA),
-             plot.background = element_rect(fill = "transparent", colour = NA))
-cbPalette <- c("#000000", "#E69F00", "#56B4E9", "#009E73", 
-               "#F0E442", "#0072B2", "#D55E00", "#CC79A7")
+library(knitcitations)
+library(nonparametricbayes) 
 ```
 
 
@@ -158,7 +150,7 @@ All parametric Bayesian estimates use the following basic parameters for the JAG
 ```{r jags-setup}
 y <- obs$x[-1] 
 N <- length(y);
-jags.data <- list("N"=N,"y"=y)
+jags.data <- list("N","y")
 n.chains <- 6
 n.iter <- 5e5
 n.burnin <- floor(10000)
@@ -241,7 +233,7 @@ parameters in the transformed space used by the MCMC.  We use logarithms
 to maintain strictly positive values of parameters where appropriate.
 
 
-```{r allen-mcmc, dependson=c("allen-model", "allen-pars", "jags-setup")}
+```{r allen-mcmc, dependson=c("allen-model", "allen-pars", "jags-setup"), results="hide"}
 jags.params=c("K","logr0","logtheta","stdQ", "stdR") # be sensible about the order here
 
 # Can give a list of starting values or a function that can create random values 
@@ -366,7 +358,7 @@ parameters in the transformed space used by the MCMC.  We use logarithms
 to maintain strictly positive values of parameters where appropriate.
 
 
-```{r ricker-mcmc, dependson="ricker-model"}
+```{r ricker-mcmc, dependson="ricker-model", results="hide"}
 # Uniform priors on standard deviation terms
 jags.params=c("K","logr0", "stdQ", "stdR")
 jags.inits <- function(){
@@ -386,7 +378,7 @@ ricker_jags <- do.call(jags.parallel,
 recompile(ricker_jags)
 ricker_jags <- do.call(autojags, 
                        list(object=ricker_jags, n.update=n.update, n.iter=n.iter, 
-                            n.thin = n.thin, progress.bar="none"))
+                            n.thin = n.thin))
 ```
 
 
@@ -449,9 +441,8 @@ paste(sprintf(
   stdR_prior_p[1], stdR_prior_p[2]),
 
   "
-  iQ <- 1 / (stdQ * stdQ);
-  iR <- 1 / (stdR * stdR);
-
+  iQ <- 1 / (stdQ * stdQ)
+  iR <- 1 / (stdR * stdR)
   r0 <- exp(logr0)
   theta <- exp(logtheta)
   K <- exp(logK)
@@ -482,25 +473,26 @@ par_priors <- list( deviance = function(x) 0 * x, logK = logK_prior,
 ```
 
 
-```{r myers-mcmc, dependson="myers-model"}
+```{r myers-mcmc, dependson="myers-model", results="hide"}
 jags.params=c("logr0", "logtheta", "logK", "stdQ", "stdR")
 jags.inits <- function(){
-  list("logr0" = log(rlnorm(1,0,.5)), 
-       "logK" = log(10 * rlnorm(1,0,.2)),
-       "logtheta" = log( 2 * rlnorm(1,0,.2)),  
-       "stdQ"= 0.2 * rlnorm(1,0,.1),
-       "stdR"= 1e-5 * rlnorm(1,0,.1),
-       "x"=y,
+  list("logr0" 		=        rnorm(1, 1, .5), 
+       "logK" 		= 10   * rnorm(1, 1, .2),
+       "logtheta" = 2    * rnorm(1, 1, .2),  
+       "stdQ"			= 0.2  * rlnorm(1, 0, .1),
+       "stdR"			= 1e-5 * rlnorm(1, 0, .1),
+       "x" 				= y,
        .RNG.name="base::Wichmann-Hill", .RNG.seed=123)
 }
 set.seed(12345)
 myers_jags <- do.call(jags.parallel, 
                       list(data=jags.data, inits=jags.inits, jags.params, 
                            n.chains=n.chains, n.iter=n.iter, n.thin=n.thin,
-                           n.burnin=n.burnin, model.file="myers.bugs")
+                           n.burnin=n.burnin, model.file="myers.bugs"))
+recompile(myers_jags)
 myers_jags <- do.call(autojags, 
                       list(object=myers_jags, n.update=n.update, n.iter=n.iter, 
-                           n.thin = n.thin, progress.bar="none"))
+                           n.thin = n.thin))
 ```
 
 Convergence diagnostics for parametric bayes