R/helloUnBBayes.R

In rommelnc/runbbayes: It provides integration to UnBBayes probabilistic framework using rJava

helloUnBBayes <- function() {
  
  ## Load Network
  file = .jnew("java/io/File", "./asia.net")
  ## .net
  io = .jnew("unbbayes/io/NetIO")
  ## .xml
  #   io = .jnew("unbbayes/io/XMLBIFIO")
  graph = io$load(file) 
  
  net = .jcast(graph, "unbbayes/prs/bn/ProbabilisticNetwork")
  attach( javaImport( "unbbayes.prs" ), pos = 2 , name = "java:unbbayes.prs.bn" )
  attach( javaImport( "unbbayes.prs.bn" ), pos = 3 , name = "java:unbbayes.prs.bn" )
  
  
  
  
  ## adding a new node manually
  newNode = new(ProbabilisticNode)
  newNode$setName("K")
  newNode$setDescription("A test node")
  newNode$appendState("State 0")
  newNode$appendState("State 1")
  auxCPT = newNode$getProbabilityFunction()
  auxCPT$addVariable(newNode)
  net$addNode(newNode)
  
  ## search for "HasVisitedAsia"
  asiaNode = .jcast(net$getNode("A"), "unbbayes/prs/bn/ProbabilisticNode")
  
  ## adding a new edge from "HasVisitedAsia" to new node
  net$addEdge(new(Edge, asiaNode, newNode))
  
  ## filling CPT of new node
  auxCPT$addValueAt(0L, .jfloat(0.99))
  auxCPT$addValueAt(1L, .jfloat(0.01))
  auxCPT$addValueAt(2L, .jfloat(0.1))
  auxCPT$addValueAt(3L, .jfloat(0.9))
  
  ## prepare the algorithm to compile network
  net = compileNetwork(net)
  
  ## print node's initial states
  prior = c()
  for (index in 1:net$getNodeCount()-1L) {
    print(index)
    node = net$getNodeAt(index)
    value = paste(node$getDescription(), ": [ ", sep="", collapse="")
    for (i in 1:node$getStatesSize()-1L) {
      value = paste(value, node$getStateAt(i), " : ", 
                    round(.jcast(node, "unbbayes/prs/bn/ProbabilisticNode")$getMarginalAt(i), 2),
                    " ", sep="", collapse="")
    }
    value = paste(value, "]", sep="", collapse="")
    prior = c(prior, value)
  }
  prior
  
  ## insert evidence (finding) to the 1st node of "net"
  findingNode = .jcast(net$getNodes()$get(0L), "unbbayes/prs/bn/ProbabilisticNode")
  findingNode$addFinding(0L) ## the 1st state is now 100%
  
  ## insert likelihood
  likelihood = .jarray(.jfloat(1:newNode$getStatesSize()))
  likelihood[[1]] = .jfloat(1)
  likelihood[[2]] = .jfloat(0.8)
  newNode$addLikeliHood(likelihood)
  
  ## propagate evidence
  net$updateEvidences()
  
  ## print updated node's states
  posterior = c()
  for (index in 1:net$getNodeCount()-1L) {
    node = net$getNodeAt(index)
    value = paste(node$getDescription(), ": [ ", sep="", collapse="")
    for (i in 1:node$getStatesSize()-1L) {
      value = paste(value, node$getStateAt(i), " : ", 
                    round(.jcast(node, "unbbayes/prs/bn/ProbabilisticNode")$getMarginalAt(i), 2),
                    " ", sep="", collapse="")
    }
    value = paste(value, "]", sep="", collapse="")
    posterior = c(posterior, value)
  }
  posterior
  
  return(rbind(prior, posterior))
}