(a) (b)
Fig. 3.43. Two C5.0 trees built for the Iris data when trails is ten.
(a) (b)
a) The C5.0 tree constructed for the seeds data. (b) The ROC curves of the C5.0
the CART model built for the seeds data.
The confusion matrix of the C5.0 model and the confusion matrix of the CART
tructed for the seeds data.
CART
C5.0
A
B
%
A
B
%
A
58
12
91.7
61
10
85.9
B
10
55
84.6
5
60
92.3
%
85.3
80.9
82.1
92.4
85.7
87.8
eeds classification
44(a) shows a C5.0 tree for the seeds data. The CART tree is
o the C5.0 tree for this data set. The tree only employed one
ng rule GrooveL<5.533, where GrooveL stands for length of
