Database Reference
In-Depth Information
Fig. 10.7 Opening the J48 tree visualization.
in Figure 10.7. Following the last action, a graphical visualization of the
J47 is presented (see Figure 10.8).
10.2.2
Building a Forest
Weka has a very rich list of ensemble algorithms that can be used to build
a decision forest. Ensemble algorithms that were specifically designed for
decision trees, such as random forest, are located in the trees section of the
hierarchical menu. Other ensemble methods such as bagging, AdaBoost and
rotation forest can be applied to any classification algorithm and therefore
are located in the meta section. Most meta learning algorithms include
one parameter for specifying the base classifier and one parameter for
Search WWH ::




Custom Search