Question

我第一次使用Rpart但却无法理解一些事情：

 str(X1)

＆＃39; data.frame＆＃39;：7329 obs。 4个变量： $ totalordered：num 34 800 36 254 564 ... $ number_of_orders：int 1 4 1 2 4 1 2 1 1 1 ... $ custom_field_classification：因子w / 5级＆＃34;教育：教育者/学生＆＃34;，..：2 2 1 2 1 1 1 2 4 3 ... $ CLTV：int 0 363 0 5 114 0 119 0 0 0 ...

>model1 <- rpart(custom_field_classification ~.,data=X1,method="class")
>model1

n = 7329 node），split，n，loss，yval，（yprob） *表示终端节点
1) root 7329 4591 Education: Educator/Student (0.37 0.24 0.12 0.23 0.043)  
  2) totalordered< 104.715 4043 1898 Education: Educator/Student (0.53 0.15 0.092 0.19 0.038) *
  3) totalordered>=104.715 3286 2151 Education: K-12 (0.18 0.35 0.15 0.27 0.051) *

>summary(model1)

Call:
rpart(formula = custom_field_classification ~ ., data = X1, method = "class")
  n= 7329 

         CP nsplit rel error    xerror        xstd
1 0.1180571      0 1.0000000 1.0000000 0.009020710
2 0.0100000      1 0.8819429 0.8832498 0.009270542

Variable importance
    totalordered number_of_orders             CLTV 
              54               26               20 

Node number 1: 7329 observations,    complexity param=0.1180571
  predicted class=Education: Educator/Student  expected loss=0.6264156  P(node) =1
    class counts:  2738  1724   866  1683   318
   probabilities: 0.374 0.235 0.118 0.230 0.043 
  left son=2 (4043 obs) right son=3 (3286 obs)
  Primary splits:
      totalordered     < 104.715 to the left,  improve=312.20560, (0 missing)
      number_of_orders < 2.5     to the left,  improve= 93.21119, (0 missing)
      CLTV             < 0.5     to the left,  improve= 63.99499, (0 missing)
  Surrogate splits:
      number_of_orders < 1.5     to the left,  agree=0.773, adj=0.493, (0 split)
      CLTV             < 0.5     to the left,  agree=0.718, adj=0.370, (0 split)

Node number 2: 4043 observations
  predicted class=Education: Educator/Student  expected loss=0.4694534  P(node) =0.5516442
    class counts:  2145   589   370   787   152
   probabilities: 0.531 0.146 0.092 0.195 0.038 

Node number 3: 3286 observations
  predicted class=Education: K-12              expected loss=0.6545953  P(node) =0.4483558
    class counts:   593  1135   496   896   166
   probabilities: 0.180 0.345 0.151 0.273 0.051

为什么我的所有因素水平都没有显示？（我有5个级别，只显示两个级别？）
为什么我只能看到总排序而不是其他参数？喜欢订单数量，CLTV？

rpart包没有显示完整的情节

0 个答案: