ID3 decision tree algorithm [8] is a well-known inductive
learning algorithm. It works on a set of examples of discrete
attribute values. Each attribute measures some feature of an object
or event. Firstly, ID3 selects an attribute for the test at the
root of the decision tree according to the classification ability.
Then, it creates a branch for each value of the selected attribute.
This procedure is recursively done for each non-terminal node
of the tree. Thus, the whole tree can be built from top to down.
Information gain rate of entropy is used as the indicator of attributes’
classification ability. C4.5 [9] and C5.0 classifiers are
developed from ID3 based on the information entropy. CART
[10] is another successful decision tree algorithm. It applies
GINI index as the indicator of attributes’ classification ability.
GINI index is similar to the entropy. There are researches [6, 4]
which focused on the speed of building decision trees. There
are also studies [7, 11] which concerned the accuracy of the
decision trees.