rpart - testing

Ref: 應用 R 語言於資料分析-從機器學習、資料探勘到巨量資料(p6-3~6-5)

1. Download R for Windows (http://cran.csie.ntu.edu.tw/)
2. Install and RUN

3. 
使用rpart套件，iris data
>library(rpart)
> data(iris)

> np = ceiling(0.1*nrow(iris))  # 
> np
[1] 15
> test.index = sample(1:nrow(iris),np)
> iris.testdata = iris[test.index,]
> iris.traindata = iris[-test.index,]
> iris.tree = rpart(Species ~ Sepal.Length + Sepal.Width + Petal.Length + Petal.Width, method="class", data=iris.traindata)
> iris.tree
n= 135 

node), split, n, loss, yval, (yprob)
      * denotes terminal node

1) root 135 89 versicolor (0.31851852 0.34074074 0.34074074)  
  2) Petal.Length< 2.45 43  0 setosa (1.00000000 0.00000000 0.00000000) *
  3) Petal.Length>=2.45 92 46 versicolor (0.00000000 0.50000000 0.50000000)  
    6) Petal.Width< 1.75 50  5 versicolor (0.00000000 0.90000000 0.10000000) *
    7) Petal.Width>=1.75 42  1 virginica (0.00000000 0.02380952 0.97619048) *
> plot(iris.tree); text(iris.tree)

>