使用 R 的决策树

Question

Given this data:

       Siblings Children Gender Survival
1         Y        Y      F        Y
2         N        Y      M        N
3         Y        Y      M        Y
4         N        N      F        N
5         Y        N      F        N
6         N        Y      F        N
7         Y        Y      M        N
8         Y        Y      M        Y
9         Y        N      F        Y
10        Y        Y      F        N
11        Y        N      M        N

When I use the function:

fit <- rpart(Survival ~ Gender + Children + Siblings,
             data = data1, method = "class")
plot(fit)

There is an error message that says:

fit is not a tree, just a root.

How can I fit and plot a decision tree?

Answer 1

When I entered your code, it also displayed the same message. I checked what it looks like inside:

> fit
n= 11 

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

1) root 11 4 N (0.6363636 0.3636364) *

It seems that this data is simply not enough for R to create something meaningful.

The fancyRpartPlot function from the rattle package works better with smaller data sets: (Taken from this answer: Decision trees in smaller datasets ):

library(rattle)
fit <- rpart(Survival ~ Gender + Children + Siblings,
             data = data, method = "class",
             control = rpart.control(minbucket=2))
fancyRpartPlot(fit, sub=NULL)

And the result is as follows: