A decision tree is a rooted, directed tree akin to a flowchart. Each internal node corresponds to a partitioning decision, and each leaf node is mapped to a class label prediction. To classify a data item, we imagine the data item to be traversing the tree, beginning at the root. Each internal node is programmedwith a splitting rule, which partitions the domain of one (or more) of the data’s attributes. Based on the splitting rule, the data item is sent forward to one of the node’s children. This testing and forwarding is repeated until the data item reaches a leaf node.