Solved – Why is CART in R not using the factor variables

cartcategorical datarrpart

I have data which looks like 

data.frame':    29969 obs. of  12 variables:
 $ X                 : int  1 2 3 4 5 6 7 8 9 10 ...
 $ Productcode       : int  102146 102146 102146 102146 102146 102146 102146 102146 102146 102146 ...
 $ Category          : Factor w/ 19 levels "[02] Livestocks",..: 1 1 1 1 1 1 1 1 1 1 ...
 $ Month             : int  1 3 4 6 7 8 9 10 11 12 ...
 $ Mode.of.operations: Factor w/ 2 levels "[1] Distribution",..: 1 1 1 1 1 1 1 1 1 1 ...
 $ sales             : int  0 0 0 0 0 0 0 0 0 0 ...
 $ profit.margin     : num  0 0 0 0 0 0 0 0 0 0 ...
 $ Name              : Factor w/ 1247 levels "★南方芝麻糊",..: 757 757 757 757 757 757 757 757 757 757 ...
 $ Packaging.content : int  32 32 32 32 32 32 32 32 32 32 ...
 $ Specifications    : Factor w/ 331 levels "0","1.018kg",..: 254 254 254 254 254 254 254 254 254 254 ...
 $ Unit              : Factor w/ 27 levels "个","代","件",..: 23 23 23 23 23 23 23 23 23 23 ...
 $ Origin            : Factor w/ 131 levels "0","409","上海",..: 82 82 82 82 82 82 82 82 82 82 ...

X     Productcode    Category     Month Mode.of.operations sales profit.margin  Name                  Packaging      Specs Unit Origin
1 1      102146 [02] Livestocks     1   [1] Distribution     0             0 洪湖松花皮蛋                32            6枚   盒   洪湖
2 2      102146 [02] Livestocks     3   [1] Distribution     0             0 洪湖松花皮蛋                32            6枚   盒   洪湖
3 3      102146 [02] Livestocks     4   [1] Distribution     0             0 洪湖松花皮蛋                32            6枚   盒   洪湖
4 4      102146 [02] Livestocks     6   [1] Distribution     0             0 洪湖松花皮蛋                32            6枚   盒   洪湖
5 5      102146 [02] Livestocks     7   [1] Distribution     0             0 洪湖松花皮蛋                32            6枚   盒   洪湖
6 6      102146 [02] Livestocks     8   [1] Distribution     0             0 洪湖松花皮蛋                32            6枚   盒   洪湖

I am trying to use classification and regression trees in R to help me understand the data. I want to visualise that which categories under what sales number or profit margin adopted which mode of operation. So I want to relate Mode of operations variable to other variables like category, sales, profit margin, origin, packaging and specs.

Based on my limited knowledge of R and Trees I tired to do it by running following command

>wushang.model = rpart(Mode.of.operations ~ sales+Packaging.content+profit.margin+Category+Origin, data=wushangtest, method ="Class", control =rpart.control(minibucket=25))

>prp(wushang.model)

But all I get is this
enter image description here

What I am doing wrong? Even if there is a little relation among variables I want R to classify and show it.

Best Answer

If "there is a little relation among variables", how can you reasonably expect R to be able to classify it?

That kind of make me think maybe you have wrong expectation about CART. Without knowing your data, a few possibilities:

  1. nothing wrong with CART, those two variables indeed are the one that can explain best the predicted variable
  2. your tree is not deep enough, if you want to see other variables (but in any case, the top portion of the tree would highlight already the most relevant variables)
  3. your data is skewed (e.g. much more of [1] Dist than [2] Rese) such that the prediction is biased.

In any case, this may not even a code issue, and may belong less in Stack Overflow than other fora e.g. Cross-Validate (for stats) or Data Science communities.

Related Question