Selecting a Subtree
13
NUMERIC_
Obs
NODE
ROLE
RANK
STAT
VALUE
CHARACTER_VALUE
1
1
PRIMARY
1
VARIABLE
.
Stores
2
1
PRIMARY
1
LABEL
.
Number of Stores
3
1
PRIMARY
1
MISSING
1.0000
4
1
PRIMARY
1
WORTH
15.5622
5
1
PRIMARY
1
BRANCHES
2.0000
6
1
PRIMARY
1
INTERVAL
10.5000
107
3
PRIMARY
1
VARIABLE
.
Product
108
3
PRIMARY
1
MISSING
1.0000
109
3
PRIMARY
1
WORTH
13.3733
110
3
PRIMARY
1
BRANCHES
2.0000
111
3
PRIMARY
1
NOMINAL
1.0000
BOOT
112
3
PRIMARY
1
NOMINAL
1.0000
SANDAL
113
3
PRIMARY
1
NOMINAL
1.0000
SLIPPER
114
3
PRIMARY
1
NOMINAL
1.0000
SPORT SHOE
115
3
PRIMARY
1
NOMINAL
2.0000
MEN’S CASUAL
116
3
PRIMARY
1
NOMINAL
2.0000
MEN’S DRESS
117
3
PRIMARY
1
NOMINAL
2.0000
WOMEN’S CASUAL
118
3
PRIMARY
1
NOMINAL
2.0000
WOMEN’S DRESS
156
6
PRIMARY
1
VARIABLE
.
Product
157
6
PRIMARY
1
MISSING
1.0000
158
6
PRIMARY
1
WORTH
7.4924
159
6
PRIMARY
1
BRANCHES
2.0000
160
6
PRIMARY
1
NOMINAL
1.0000
BOOT
161
6
PRIMARY
1
NOMINAL
1.0000
SANDAL
162
6
PRIMARY
1
NOMINAL
1.0000
SPORT SHOE
163
6
PRIMARY
1
NOMINAL
2.0000
SLIPPER
171
7
PRIMARY
1
VARIABLE
.
Product
172
7
PRIMARY
1
MISSING
2.0000
173
7
PRIMARY
1
WORTH
1.2574
174
7
PRIMARY
1
BRANCHES
2.0000
175
7
PRIMARY
1
NOMINAL
1.0000
MEN’S CASUAL
176
7
PRIMARY
1
NOMINAL
2.0000
MEN’S DRESS
177
7
PRIMARY
1
NOMINAL
2.0000
WOMEN’S CASUAL
178
7
PRIMARY
1
NOMINAL
2.0000
WOMEN’S DRESS
Figure 5.
Primary Splitting Rules
Node 1 represents the entire training data. The rows in the listing for which col-
umn NODE is 1 describe the rule for partitioning node 1. The STAT column deter-
mines the type of information in the row. Rows with STAT equal to ‘VARIABLE’
and ‘LABEL’ identify the name and label of the input variable used by the splitting
rule. In this example, the splitting rule is based on Number of Stores, creates two
branches (STAT equals ‘BRANCHES’), and assigns data observations with Stores
< 10.5 (STAT equals ‘INTERVAL’) to the first branch.
The splitting rule for node 3 assigns to the first branch observations in which Product
equals ‘BOOT’, ‘SANDAL’, ‘SLIPPER’, or ‘SPORT SHOE’. In node 7, ‘MEN’S
CASUAL’ are assigned to the first branch, and ‘MEN’S DRESS’, ‘WOMEN’S
CASUAL’, and ‘WOMEN’S DRESS’ to the second branch.
