Cost-Sensitive Learning for decision trees
Decision trees are binary trees that use conditional decision-making to predict the class of the samples. Every tree node represents a set of samples corresponding to a chain of conditional statements based on the features. We divide the node into two children based on a feature and a threshold value. Imagine a set of students with height, weight, age, class, and location. We can divide the set into two parts according to the features of age and with a threshold of 8. Now, all the students with ages less than 8 will go into the left child, and all those with ages greater than or equal to 8 will go into the right child.
This way, we can create a tree by successively choosing features and threshold values. Every leaf node of the tree will contain nodes from only one class, respectively.
A question often arises during the construction of a decision tree: “Which feature and threshold pair should be selected to partition the set of...
