Question

我正在尝试在简单的数据集上执行朴素贝叶斯分类器。我拥有的三个变量是weight（连续），BP（连续）和disease（二分）。

当我为朴素贝叶斯编写命令时，一些结果给了我一个大于一的概率。我也通过＆＃39; e1071＆＃39;和＆＃39; klaR＆＃39;。

请参阅下面的代码：

> install.packages("e1071")

> library(e1071)

> mydata$disease<-as.factor(mydata$disease)

> classifier<- naiveBayes(disease ~ weight + BP, mydata, laplace = 0, subset, na.action = na.pass)

> Please see my results below,

> A-priori probabilities:

> Y

>    0    1 

> 0.47 0.53 

> Conditional probabilities:

>    weight

> Y        [,1]     [,2]

>   0  69.10638 27.22869

>   1 131.22642 39.47377

>    BP

> Y       [,1]     [,2]

>   0 44.78723 21.73350

>   1 35.81132 13.55623

如上所示，其中一个概率是44.78723。那是对的吗？我也尝试了klaR，它给了我非常相似的结果。帮助

Answer 1

升级评论：

输出给出了类变量的每个级别的正态分布的参数（平均值和标准差）。来自?naiveBayes帮助：

For each numeric variable, a table giving, for each target class, mean and 
standard deviation of the (sub-)variable)

使用iris数据集

运行一个小示例来查看此内容

library(e1071)

# load iris dataset and set some values to missing
data(iris)
iris$Sepal.Length[1] <- NA
iris$Petal.Width[2] <- NA
iris$Species[3] <- NA

# run naive Bayes model
(m <- naiveBayes(Species ~ Sepal.Length + Petal.Width , data = iris, na.action=na.omit))

这会产生输出

# Naive Bayes Classifier for Discrete Predictors
# 
# Call:
#   naiveBayes.default(x = X, y = Y, laplace = laplace)
# 
# A-priori probabilities:
#   Y
# setosa versicolor  virginica 
# 0.3197279  0.3401361  0.3401361 
# 
# Conditional probabilities:
#   Sepal.Length
# Y                [,1]      [,2]
# setosa     5.012766 0.3603241
# versicolor 5.936000 0.5161711
# virginica  6.588000 0.6358796
# 
# Petal.Width
# Y                 [,1]      [,2]
# setosa     0.2489362 0.1080908
# versicolor 1.3260000 0.1977527
# virginica  2.0260000 0.2746501

检查表格是否给出了平均值和st。每个Species

级别的每个连续变量的dev

aggregate(cbind(Sepal.Length, Petal.Width) ~ Species, data=iris, 
                        function(i) c(mean(i), sd(i)))

# Species Sepal.Length.1 Sepal.Length.2 Petal.Width.1 Petal.Width.2
# 1     setosa      5.0127660      0.3603241     0.2489362     0.1080908
# 2 versicolor      5.9360000      0.5161711     1.3260000     0.1977527
# 3  virginica      6.5880000      0.6358796     2.0260000     0.2746501

naiveBayes命令给出大于零的概率

1 个答案: