數據分析與挖掘 - R語言：貝葉斯分類算法（案例二）

接着案例一，我們再使用另一種方法實例一個案例

 

直接上代碼：

#!/usr/bin/Rscript

library(plyr)
library(reshape2)

#1、根據訓練集創建朴素貝葉斯分類器
#1.1、生成類別的概率

##計算訓練集合D中類別出現的概率，即P{c_i}
##輸入：trainData 訓練集，類型為數據框
##      strClassName 指明訓練集中名稱為    strClassName列為分類結果
##輸出：數據框，P{c_i}的集合，類別名稱|概率（列名為 prob）
class_prob <- function(trainData, strClassName){
    #訓練集樣本數
    #nrow返回行數
    length.train <- nrow(trainData)
    dTemp <- ddply(trainData, strClassName, "nrow")
    dTemp <- ddply(dTemp, strClassName, mutate, prob = nrow/length.train)
    dTemp[,-2]
}

##1.2、生成每個類別下，特征取不同值的概率
##計算訓練集合D中,生成每個類別下，特征取不同值的概率，即P{fi|c_i}
##輸入：trainData 訓練集，類型為數據框
##      strClassName 指明訓練集中名稱為strClassName列為分類結果，其余的全部列認為是特征值
##輸出：數據框，P{fi|c_i}的集合，類別名稱|特征名稱|特征取值|概率（列名為 prob）
feature_class_prob <- function(trainData, strClassName){
    # 橫表轉換為縱表
    data.melt <- melt(trainData,id=c(strClassName))
    # 統計頻數
    aa <- ddply(data.melt, c(strClassName,"variable","value"), "nrow")
    # 計算概率
    bb <- ddply(aa, c(strClassName,"variable"), mutate, sum = sum(nrow), prob = nrow/sum)
    # 增加列名
    colnames(bb) <- c("class.name",
                    "feature.name",
                    "feature.value",
                    "feature.nrow",
                    "feature.sum",
                    "prob")
    # 返回結果
    bb[,c(1,2,3,6)]
}

## 以上創建完朴素貝葉斯分類器


## 2、使用生成的朴素貝葉斯分類器進行預測
##使用生成的朴素貝葉斯分類器進行預測P{fi|c_i}
##輸入：oneObs 數據框，待預測的樣本，格式為 特征名稱|特征值
##      pc 數據框，訓練集合D中類別出現的概率，即P{c_i}  類別名稱|概率
##      pfc 數據框，每個類別下，特征取不同值的概率，即P{fi|c_i}
##                  類別名稱|特征名稱|特征值|概率
##輸出：數據框，待預測樣本的分類對每個類別的概率，類別名稱|后驗概率（列名為 prob）
pre_class <- function(oneObs, pc,pfc){
    colnames(oneObs) <- c("feature.name", "feature.value")
    colnames(pc) <- c("class.name","prob")
    colnames(pfc) <- c("class.name","feature.name","feature.value","prob")

    # 取出特征的取值的條件概率
    feature.all <- join(oneObs,pfc,by=c("feature.name","feature.value"),type="inner")
    # 取出特征取值的條件概率連乘
    feature.prob <- ddply(feature.all,.(class.name),summarize,prob_fea=prod(prob))  #prod為連乘函數

    #取出類別的概率
    class.all <- join(feature.prob,pc,by="class.name",type="inner")
    #輸出結果
    ddply(class.all,.(class.name),mutate,pre_prob=prob_fea*prob)[,c(1,4)]
}


##3、數據測試
##用上面蘋果的數據作為例子進行測試
#訓練集
train.apple <-data.frame(
    size=c("大","小","大","大","小","小"),
    weight=c("輕","重","輕","輕","重","輕"),
    color=c("紅","紅","紅","綠","紅","綠"),
    taste=c("good","good","bad","bad","bad","good")
)
#待預測樣本
oneObs<-data.frame(
    feature.name =c("size", "weight", "color"),
    feature.value =c("大","重","紅")
)

#預測分類
pc <- class_prob(train.apple,"taste")
pfc <- feature_class_prob(train.apple,"taste")
pre_class(oneObs, pc, pfc)

預測結果為：

class.name pre_prob
1 bad 0.07407407
2 good 0.03703704

可見該蘋果的口味為：bad

 

*********************************************這里是分割線****************************************************

我們使用這個方法再預測一下案例一中的數據集。

#數據集樣本
data <- data.frame(c("sunny","hot","high","weak","no",  
                 "sunny","hot","high","strong","no",  
                 "overcast","hot","high","weak","yes",  
                 "rain","mild","high","weak","yes",  
                 "rain","cool","normal","weak","yes",  
                 "rain","cool","normal","strong","no",  
                 "overcast","cool","normal","strong","yes",  
                 "sunny","mild","high","weak","no",  
                 "sunny","cool","normal","weak","yes",  
                 "rain","mild","normal","weak","yes",  
                 "sunny","mild","normal","strong","yes",  
                 "overcast","mild","high","strong","yes",  
                 "overcast","hot","normal","weak","yes",  
                 "rain","mild","high","strong","no"), 
                 byrow = TRUE,
                 dimnames = list(day = c(),condition = c("outlook","temperature","humidity","wind","playtennis")), 
                 nrow=14, 
                 ncol=5);  

#待預測樣本
ddata<-data.frame(
    feature.name =c("outlook", "temperature","humidity","wind"),
    feature.value =c("overcast","mild","normal","weak")
)


#預測分類
pc <- class_prob(data,"playtennis")
pfc <- feature_class_prob(data,"playtennis")
pre_class(ddata, pc, pfc)

預測結果為：

class.name   pre_prob
1         no 0.02666667
2        yes 0.13168724

預測結果為：yes，可見與案例一的結果一樣。