R-数据挖掘-贝叶斯分类（五）

海林老师《数据挖掘》（韩佳炜书）课程作业系列

要求：自己写R/Python代码、函数实现一系列算法

其他参见：

R-数据挖掘-求混合型数据对象距离（一）

R-数据挖掘-主成分分析PCA（二）

R-数据挖掘-关联规则（三）

R-数据挖掘-决策树ID3（四）

R-数据挖掘-贝叶斯分类（五）

R-数据挖掘-聚类Kmeans（六）

R-数据挖掘-聚类DBSCAN（七）

全文逻辑：

1. 分析
2. 算法/函数
3. 测试数据
4. 测试代码
5. 测试结果（截图）

分析：

#输入训练集(数据框,要求其最后一列为类别)和测试集(向量)
#输出print在每个类的概率值
#返回训练集的类别

 

算法实现（编写函数）：

bys<-function(data,test){
#输出分类名
#默认最后一列为分类
  get_classname<-function(data){
    classes=data[,ncol(data)]
    return(names(table(classes)))
  }
  #根据分类切割数据
  #classname为一个字符串
  split_data<-function(data,classname){
    data_fen=subset(data,data[,ncol(data)]==classname)
    data_fen=droplevels(data_fen)
    return(data_fen)
  }
  #计算某个分类比例
  #传入分类名
  p_class<-function(data,classname){
    return(nrow(data[data$class==classname,])/nrow(data))
  }
  
  #计算 某个条件属性，某个分类属性的概率 即求P(x1|C1)
  ###attr_index属性列序号
  ###attr_value属性的具体值
  ###classname某个分类名
  p_attr_class<-function(data_fen,attr_index,attr_value){
    count_classname=nrow(data_fen)
    count_attr=nrow(data_fen[data_fen[,attr_index]==attr_value,])
    return(count_attr/count_classname)
  }
  
  #输出结果函数（为test选择最合适的类别）
  p<-function(data,test){
    classname=get_classname(data)
    print(classname)
    p_cs=c()
    for (i in 1:length(classname)) {
      p_class_i=p_class(data,classname[i])
      data_fen=split_data(data,classname[i])
      p_x_ci=1
      for (j in 1:length(test)) {
        p_x_ci=p_x_ci* p_attr_class(data_fen,j,test[j])
      }
      p_cs[i]=p_class_i*p_x_ci
    }
    print(p_cs)
    ##选择概率最大的
    index=which(rank(-as.numeric(p_cs),ties.method="first")==1)
    return(classname[index])
  }
  return(p(data,test))
}

数据测试：

测试数据：

如下

########################################训练集
age=c("youth","youth","youth","youth","youth","middle_aged","middle_aged","middle_aged","middle_aged","senior","senior","senior","senior","senior")
income=c("high","high","medium","low","medium","high","low","medium","high","medium","low","low","medium","medium")
student=c("no","no","no","yes","yes","no","yes","no","yes","no","yes","yes","yes","no")
credit_rating=c("fair","excellent","fair","fair","excellent","fair","excellent","excellent","fair","fair","fair","excellent","fair","excellent")
class=c("no","no","no","yes","yes","yes","yes","yes","yes","yes","yes","no","yes","no")
train<-data.frame(age,income,student,credit_rating,class,stringsAsFactors = F)
#################################################测试集
#测试集：不考虑缺失值情况
test<-c("youth","medium","yes","fair")

xx=bys(train,test)

测试结果：