A Walk-through-guide for Using Decision Trees Software - Data Mining with Decision Trees: Theory and Applications - page 153

Fig. 10.1 Weka preprocessing.

Figure 10.2 illustrates the preprocess windows after loading the Iris dataset.
Note that on the left part of the screen, we can see the list of attributes
in the Iris dataset is presented. The right side shows the properties of the
selected attribute and how the selected attribute and the target attribute
are co-distributed. In the preprocess, it is possible to edit the dataset
by clicking the Edit button or run a filtering procedure (such as feature
selection) on the raw dataset. This is done by choosing a filter and clicking
on Apply.

10.2.1 Training a Classification Tree

Weka implements several decision tree induction algorithms. The most
frequently used algorithm is the J48 which is a variation of the well-known
C4.5 algorithm development. While many consider J48 identical to C45,
it should be noted that there are some differences. A comparison study
performed by Moore et al. (2009) reveals that J48 consistently performs
worse than C4.5 on the datasets that were evaluated.