#loadData
loadData <- function(){
  dataSet = matrix(c(1,1,0,"yes",1,1,1,"yes",1,0,1,"no",0,1,0,"no",0,1,1,"no"),byrow = T, nrow =5)
  colnames(dataSet) = c("a","b","c","labels")
  return (dataSet)
}

#computing ShannonEnt 實現夏農熵
calShannonEnt <- function(dataSet){
  numEntries = nrow(dataSet)
  labels = levels(factor(dataSet[,"labels"]))
  labelCount = NULL
  labelCount[labels] = rep(0,length(labels))
  for(i in 1:numEntries){
    if(dataSet[i,"labels"] %in% labels){
      temp = dataSet[i,"labels"]
      labelCount[temp] = labelCount[temp] + 1
    }
  }
  t = NULL
  shannonEnt = 0
  for(i in  1:length(labelCount)){
    t[i]= labelCount[i]*1.0 / numEntries
    shannonEnt = -t[i]*log2(t[i]) + shannonEnt
  }
#  labelCount = as.numeric(labelCount)
  return (shannonEnt)
}

#劃分資料集 split DataSet
splitDataSet <- function(dataSet,axis,value){
  retDataSet = NULL
  for(i in 1:nrow(dataSet)){
    if(dataSet[i,axis] == value){
      tempDataSet = dataSet[i,]
      retDataSet = rbind(retDataSet,tempDataSet)
    }
  }
  rownames(retDataSet) = NULL
  return (retDataSet)
}

#choose Best Feature To Split 選擇最好的資料集
chooseBestFeatureToSplita <- function(dataSet){
  numFeatures = ncol(dataSet) - 1
  baseEntropy = calShannonEnt(dataSet)
  #最大資訊增益
  bestInfoGain = 0.0
  bestFeature = -1 
  for(i in  1: numFeatures){
    featureLabels = levels(factor(dataSet[,i]))
   # featureLabels = as.numeric(featureLabels)
    newEntropy = 0.0
    for( j in 1:length(featureLabels)){
      subDataSet = splitDataSet(dataSet,i,featureLabels[j])
      prob = length(subDataSet[,1])*1.0/nrow(dataSet)
      newEntropy = newEntropy + prob*calShannonEnt(subDataSet)
    }
    infoGain = baseEntropy - newEntropy
    if(infoGain > bestInfoGain){
      bestInfoGain = infoGain
      bestFeature = i
    }
  }
  return (bestFeature)
} 

#最終判斷屬於哪一類的條件
majorityCnt <- function(classList){
  classCount = NULL
  count = as.numeric(table(classList))
  majorityList = levels(as.factor(classList))
  if(length(count) == 1){
    return (majorityList[1])
  }else{
    f = max(count)
    return (majorityList[which(count == f)][1])
  }
}

#little trick for R Language,judgement whether belongs to one classify
trick <- function(classList){
  count = as.numeric(table(classList))
  if(length(count) == 1){
    return (TRUE)
  }else
    return (FALSE)
}

#遞迴建立生成樹
creatTree <- function(dataSet){########temp是所有的需要分割的特徵
  decision_tree = list()
  classList = dataSet[,"labels"]
    #判斷是否屬於同一類
  if(trick(classList))
    return (rbind(decision_tree,classList[1]))
  #是否在矩陣中只剩Label標籤了，若只剩Label標籤，則都分完了
  if(ncol(dataSet) == 1){
    decision_tree = rbind(decision_tree,majorityCnt(classList))
    return (decision_tree)
  }
  #資料集中只剩下一個元素了 
  ########選擇bestFeature作為分割屬性
  bestFeature = chooseBestFeatureToSplita(dataSet)
  ########，接下來就需要把它刪除
  labelFeature = colnames(dataSet)[bestFeature]
  #DIECISION TREE
  decision_tree = rbind(decision_tree,labelFeature)
  #選中了那個標籤作為此次分類標籤 
  t = dataSet[,bestFeature]
  temp_tree = data.frame()
  for(j in 1:length(levels(as.factor(t)))){
    #這個標籤的兩個屬性，比如“yes”，“no”，所屬的資料集
    dataSet = splitDataSet(dataSet,bestFeature,levels(as.factor(t))[j])
    dataSet[,-bestFeature]
    #遞迴呼叫這個函式
    temp_tree = creatTree(dataSet)
    decision_tree = rbind(decision_tree,temp_tree)
  }
  
  return (decision_tree)
}
 

#載入資料
dataSet = loadData()

#建立決策樹
t = creatTree(dataSet)