Merge pull request #106 from nutterb/current-devel

Current devel

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: HydeNet
 Type: Package
 Title: Hybrid Bayesian Networks Using R and JAGS
-Version: 0.10.5
+Version: 0.10.8-01
 Author: Jarrod E. Dalton <[email protected]> and Benjamin Nutter
     <[email protected]>
 Maintainer: Benjamin Nutter <[email protected]>
@@ -37,6 +37,7 @@ Imports:
     utils
 Suggests:
     knitr,
+    RCurl,
     survival,
     testthat
 VignetteBuilder: knitr
@@ -45,4 +46,4 @@ LazyLoad: yes
 LazyData: true
 URL: https://github.com/nutterb/HydeNet,
 BugReports: https://github.com/nutterb/HydeNet/issues
-RoxygenNote: 5.0.1
+RoxygenNote: 6.0.1

NAMESPACE

@@ -15,7 +15,7 @@ S3method(modelToNode,survreg)
 S3method(modelToNode,xtabs)
 S3method(plot,HydeNetwork)
 S3method(print,HydeNetwork)
-S3method(print,HydePosterior)
+S3method(print,HydeSim)
 S3method(print,cpt)
 S3method(summary,HydeNetwork)
 S3method(update,HydeNetwork)
@@ -30,7 +30,9 @@ export("%>%")
 export(HydeNetwork)
 export(HydePlotOptions)
 export(HydePosterior)
+export(HydeSim)
 export(bindPosterior)
+export(bindSim)
 export(compileDecisionModel)
 export(compileJagsModel)
 export(cpt)
@@ -53,7 +55,6 @@ export(writeNetworkModel)
 import(nnet)
 importFrom(dplyr,bind_rows)
 importFrom(dplyr,group_by_)
-importFrom(dplyr,mutate)
 importFrom(dplyr,summarise)
 importFrom(gRbase,dag)
 importFrom(gRbase,graphNEL2adjMAT)
@@ -65,7 +66,6 @@ importFrom(stats,as.formula)
 importFrom(stats,coef)
 importFrom(stats,terms)
 importFrom(stats,update)
-importFrom(stringr,perl)
 importFrom(stringr,str_extract)
 importFrom(stringr,str_extract_all)
 importFrom(stringr,str_split_fixed)

R/Hyde-package.R

@@ -7,7 +7,7 @@
 #' probability distributions, where each node represents a random variable and 
 #' each edge represents conditionality. The full joint distribution is 
 #' therefore factorized as a product of conditional densities, where each node 
-#' is assumed to be independent of its non-desendents given information on its 
+#' is assumed to be independent of its non-descendants given information on its 
 #' parent nodes. Since exact, closed-form algorithms are computationally 
 #' burdensome for inference within hybrid networks that contain a combination 
 #' of continuous and discrete nodes, particle-based approximation techniques 

R/HydeNetwork.R

@@ -44,7 +44,7 @@
 #'       to fitter functions.
 #'   \item \code{nodeParams} A named list.  Each element is a vector of parameters 
 #'       that will be expected by JAGS.
-#'   \item \code{fromData} A named list with the logical value of wheather parameters
+#'   \item \code{fromData} A named list with the logical value of whether parameters
 #'       should be estimated from the data.
 #'   \item \code{nodeData} A named list with the data for each node.  If a node's 
 #'       entry in \code{fromData} is \code{TRUE} and \code{nodeData} is \code{NULL},
@@ -200,7 +200,7 @@ HydeNetwork.formula <- function(nodes, data=NULL, ...)
                                       FUN = is.factor, 
                                       FUN.VALUE = logical(1))]
     factorLevels[factor_vars] <- 
-      lapply(X = data[, factor_vars, drop = FALSE],
+      lapply(X = data[factor_vars],
              FUN = levels)
   }
 
@@ -354,7 +354,7 @@ HydeNetwork_nodeFormula <- function(x, parents, data, fromData)
 {
   if (is.null(parents[[x]]))
   {
-    if (fromData[[names(parents)[x]]] & !is.numeric(data[, names(parents)[x]]))
+    if (fromData[[names(parents)[x]]] & !is.numeric(data[[names(parents)[x]]]))
     {
       f <- paste("~ ", names(parents)[x])
     }
@@ -381,24 +381,24 @@ HydeNetwork_nodeFitter <- function(node_name, data, parents)
   {
     return(NULL)
   }
-  else if (is.numeric(data[, node_name])) 
+  else if (is.numeric(data[[node_name]])) 
   {
     return("lm")
   }
-  else if (is.factor(data[, node_name]) & is.null(parents[[node_name]]))
+  else if (is.factor(data[[node_name]]) & is.null(parents[[node_name]]))
   {
     return("xtabs")
   }
-  else if (is.factor(data[, node_name]) & 
-           all(vapply(parents[[node_name]], function(p) is.factor(data[, p]), logical(1))))
+  else if (is.factor(data[[node_name]]) & 
+           all(vapply(parents[[node_name]], function(p) is.factor(data[[p]]), logical(1))))
   {
     return("cpt")
   }
-  else if (is.factor(data[, node_name]) & nlevels(data[, node_name]) == 2)
+  else if (is.factor(data[[node_name]]) & nlevels(data[[node_name]]) == 2)
   {
     return("glm")
   }
-  else if (is.factor(data[, node_name]) & nlevels(data[, node_name]) > 2)
+  else if (is.factor(data[[node_name]]) & nlevels(data[[node_name]]) > 2)
   {
     return("multinom")
   }
@@ -417,10 +417,10 @@ HydeNetwork_nodeType <- function(node_name, data, parents, nodeFitter)
   if (node_name %in% names(data))
   {
     if ((is.null(parents[[node_name]]) && 
-         !is.numeric(data[, node_name])) || 
+         !is.numeric(data[[node_name]])) || 
         (!is.null(parents[[node_name]]) && 
-         !is.numeric(data[, node_name]) && 
-         nlevels(data[, node_name]) > 2))
+         !is.numeric(data[[node_name]]) && 
+         nlevels(data[[node_name]]) > 2))
     {
       return('dcat')
     }
@@ -429,10 +429,10 @@ HydeNetwork_nodeType <- function(node_name, data, parents, nodeFitter)
       return('dcat')
     }
     else if ((is.null(parents[[node_name]]) && 
-              !is.numeric(data[, node_name])) || 
+              !is.numeric(data[[node_name]])) || 
              (!is.null(parents[[node_name]]) && 
-              !is.numeric(data[, node_name]) && 
-              nlevels(data[, node_name]) == 2))
+              !is.numeric(data[[node_name]]) && 
+              nlevels(data[[node_name]]) == 2))
     {
       return('dbern')
     }

R/HydePosterior.R

@@ -1,8 +1,8 @@
-#' @name HydePosterior
-#' @export HydePosterior
+#' @name HydeSim
+#' @export HydeSim
 #' 
-#' @title Posterior Distributions of a Decision Network
-#' @description The posterior distributions of the decision network can be 
+#' @title Simulated Distributions of a Decision Network
+#' @description The simulated distributions of the decision network can be 
 #'   evaluated to determine the probabilistic outcomes based on the decision
 #'   inputs in the model as well as subject specific factors.
 #'   
@@ -18,7 +18,7 @@
 #'   forced into \code{variable.names} if not already provided.  This is 
 #'   recommended, especially if you will be binding multiple JAGS runs 
 #'   together.
-#' @param bind Logical. If \code{TRUE}, posterior distributions will be bound into 
+#' @param bind Logical. If \code{TRUE}, simulated distributions will be bound into 
 #'   a single data frame.  If \code{FALSE}, the standard output from \code{rjags}
 #'   is returned.
 #'   
@@ -27,16 +27,12 @@
 #'   the rows of the policy/decision matrix given in the \code{data} argument 
 #'   of \code{compileJagsNetwork}.
 #'   
-#' @return A list of class \code{HydePosterior} with elements \code{codas} 
+#' @return A list of class \code{HydeSim} with elements \code{codas} 
 #'   (the MCMC matrices from \code{coda.samples}), \code{observed} (the values
 #'   of the variables that were observed), \code{dag} (the dag object for 
 #'   convenience in displaying the network), and \code{factorRef} (giving the
 #'   mappings of factor levels to factor variables).  
 #'   
-#'   The only rationale for giving this object its own class was because it 
-#'   produces an enormous amount of material to be printed.  A distinct 
-#'   \code{print} method has been written for this object.
-#'   
 #' @author Jarrod Dalton and Benjamin Nutter
 #' 
 #' @examples
@@ -53,20 +49,20 @@
 #' compiledNet <- compileJagsModel(Net, n.chains=5)
 #' 
 #' #* Generate the posterior distribution
-#' Posterior <- HydePosterior(compiledNet, 
+#' Posterior <- HydeSim(compiledNet, 
 #'                            variable.names = c("d.dimer", "death"), 
 #'                            n.iter = 1000)
 #' 
 #' #* Posterior Distributions for a Decision Model
 #' Net <- setDecisionNodes(Net, angio, treat)
 #' decisionNet <- compileDecisionModel(Net, n.chains=5)
-#' decisionsPost <- HydePosterior(decisionNet, 
+#' decisionsPost <- HydeSim(decisionNet, 
 #'                                variable.names = c("d.dimer", "death"),
 #'                                n.iter = 1000)
 #' 
 #' 
 
-HydePosterior <- function(cHN, variable.names, n.iter, thin=1, ...,
+HydeSim <- function(cHN, variable.names, n.iter, thin=1, ...,
                           monitor_observed=TRUE, bind=TRUE)
 {
   if (monitor_observed)
@@ -121,14 +117,23 @@ HydePosterior <- function(cHN, variable.names, n.iter, thin=1, ...,
 
 
 
-  class(HydePost) <- "HydePosterior"
+  class(HydePost) <- "HydeSim"
   if (bind)
   {
-    bindPosterior(HydePost)
+    bindSim(HydePost)
   }
   else 
   {
     HydePost
   }
 
 }
+
+#' @rdname HydeSim
+#' @export
+
+HydePosterior <- function(...)
+{
+  message("`HydePoseterior` has been deprecated and replaced by `HydeSim`")
+  HydeSim(...)
+}

R/HydeUtilities.R

@@ -1,12 +1,10 @@
 #' @name HydeUtilities
-#' @importFrom dplyr group_by_
-#' @importFrom dplyr mutate
 #' @importFrom magrittr %>%
 #' @export %>%
 #' 
 #' @title Hyde Network Utility Functions 
 #' @description The functions described below are unexported functions that 
-#'   are used internally by \code{HydeNet} to prepare modify network objects
+#'   are used internally by \code{HydeNet} to prepare and modify network objects
 #'   and prepare JAGS code.
 #'   
 #' @details 
@@ -25,7 +23,7 @@
 #'   
 #'   \code{matchLevelNumber}: Assigns the correct numeric value of a level to 
 #'   a factor variable in a model.  This is called from within 
-#'   \code{makeJagsRead}.
+#'   \code{makeJagsReady}.
 #'   
 #'   \code{matchVars}: Given a list of existing node names, the terms of a formula
 #'   are matched to the node names.  This allows functions to be used in 
@@ -89,8 +87,6 @@ termName <- function(term, reg)
 }
 
 #' @rdname HydeUtilities
-#' @importFrom stringr perl
-#' @importFrom stringr str_extract
 #' 
 #' @param node Character string indicating a node in a network
 #' @param network A Hyde Network Object

R/PolicyMatrix.R

@@ -29,19 +29,19 @@
 #'   enforced.  Thus, it is possible to include numeric values in a decision 
 #'   matrix.
 #'   
-#'   Policy matrices can be passed to \code{HydePosterior} to run posterior 
+#'   Policy matrices can be passed to \code{HydeSim} to run posterior 
 #'   distributions on each row of the policy matrix.  There is nothing 
 #'   particularly special about the policy matrices returned by 
 #'   \code{policyMatrix}; they are simply data frame that require names drawn
 #'   from the nodes in the network.  Any data frame can be passed to 
-#'   \code{HydePosterior} and a check is done there to confirm all of the 
+#'   \code{HydeSim} and a check is done there to confirm all of the 
 #'   column names match a node in the network.
 #'   
 #'   Whenever a node is identified as a deterministic node, its policy values
 #'   are forced to \code{NULL}, regardless of what the user has specified.
 #'   
 #' @return Returns a data frame built by \code{expand.grid} and intended to be
-#'   used with \code{HydePosterior}.
+#'   used with \code{HydeSim}.
 #'   
 #' @author Jarrod Dalton and Benjamin Nutter
 #'