Information Technology Reference
In-Depth Information
a.
0123456789
OHOWHbbabbaabaaa07936
01234567890
C = {95, 91, 94, 69, 89, 90, 78, 90, 69, 94}
b.
OUTLOOK
sunny
overcast
rainy
HUMIDITY
OUTLOOK
WINDY
d
90
>
90
sunny
overcast
rainy
true
false
HUMIDITY
No
No
Ye s
No
No
Yes
d
>
78
78
Ye s
No
Figure 9.14.
Perfect solution to the play tennis problem created in generation 6
(chromosome 9).
a)
The chromosome of the individual with its random numerical
constants.
b)
The corresponding decision tree. It classifies correctly all the 14
sample cases presented in Table 9.3.
splitting of the data under HUMIDITY, resulting in a perfect solution to the
problem at hand.
Let's now see how the decision trees of gene expression programming
perform on more difficult real-world applications.
9.3 Solving Problems with GEP Decision Trees
In this section we are going to put the decision trees of gene expression
programming to the test by solving four complex real-world problems. The
first is the already familiar breast cancer problem with nine numeric attributes
and two different outcomes: benign or malignant. The second is again the
iris dataset, also with numeric attributes (four, as you may recall) and three
different classes (
Iris setosa
,
Iris versicolor
, and
Iris virginica
). The third is
