Decision Tree Learning and Instances Classification
After loading a data file, the next step is to apply the decision tree learning:
After the successfull learning, two tabs are enabled (if they weren't): Inference and Verbosity. The first represents the created tree, while the second presents data about calculated gains and intermediate instance sets.
The tree represented on Inference tab can be used to classify new instances. The nodes that aren't above leaves represents a point of decision. When the user clicks on one of these nodes, the decisions possible from this point (the values of an attribute) are shown on the right. When one of these decisions are taken, the node relative to the decision is selected. If this node isn't above a leaf, the relative decisions are shown on the right. Else, the leaf classification is shown. So, the user can click on a tree node (the root is a good choice) and asks the questions on the right, what represents the attributes values of the new instance, until the classification appears in red. The images above presents the two nodes situations.