A Predictive Approach for Selecting Suitable Computing Nodes in Grid Environment by Using Data Mining Technique - Advances in Computational Science and Engineering - page 193

• Decision trees are simple to understand and interpret: At each internal node, a

decision is based upon just one predictor variable and this makes it easy to follow.
For example, to explain particular classification one need only look at the series of
simple decisions that led to it. The final tree model can in-fact be cast into a set of
rules one can follow to classify a given case.

• Decision tree methods have a built-in feature selection method that makes them

immune to the presence of useless variables.

• Tree models are very adept at revealing complex interactions between variables.

Each branch of a tree can contain different combinations of variables and the same
variable can appear more than once in different parts of the tree.

• Is robust, perform well with large data in a short time. Large amounts of data can

be analyzed using personal computers in a time short enough to enable stake-
holders to take decisions based on its analysis.

• Is a white box model. If a given situation is observable in a model the explanation

for the condition is easily explained by Boolean logic. An example of a black box
model is an artificial neural network since the explanation for the results is exces-
sively complex to be comprehended.

• It is possible to validate a model using statistical tests. That makes it possible to

account for the reliability of the model.

• Data preparation for a decision tree is basic or unnecessary. Other techniques often

require data normalization, dummy variables need to be created and blank values to
be removed.

• And other advantage.

However general decision tree always has a deterministic result, and therefore this
feature is not good in some application. Fuzzy decision Tree (FDT) is the generali-
zation of decision tree in fuzzy environment. The knowledge represented by fuzzy
decision tree is closer to the human classification [10]. FDT is combining symbolic
decision trees with approximate reasoning offered by fuzzy representation. The
intent is to exploit complementary advantages of both: popularity in applications to
learning from examples and high knowledge comprehensibility of decision trees,
ability to deal with inexact and uncertain information of fuzzy representation. FDT
has some advantage as following [19].

1. FDT is lower than DT in train accuracy while higher in test accuracy clearly, which
means that the generalization ability of FDT is better and FDT can describe the
character of customer data better.
2. The tree build by FDT is smaller and the leaf number is less, which means that
created rule is easy to understand.
3. The speed of FDT is more rapid than DT, which fits to handle huge database, such
as customer database in corporations.

In our approach we used a Fuzzy decision tree (FDT). Fuzzy decision trees provide
away to manipulate fuzzy information and continuous input/output models while
maintaining the interpretability and effectiveness of classical decision trees.