CART toy example.

<p>Classification And Regression Trees (CART) are binary decision trees, attempting to classify a pattern by selecting from a large number of variables the most important ones in determining the outcome variable. A decision tree consists of nodes and leaves, with each leaf denoting a class. In a binary tree, by convention if the answer to a question is “yes”, then the left branch is selected and the same question may appear in more than one places in the tree. For example, attributes such as “Unprotected sexual activity”, “More sexual intercourses”, “Blood transfusions (before 1985)” and “Being vaccinated against hepatitis A and/or hepatitis B” can be used to classify people as “L”, low risk of getting infected with HIV and “H”, high risk of getting infected with HIV. Classes (low and high risk of getting infected with HIV) are the outputs of the tree. Attributes (unprotected sexual activity, more sexual intercourses, possible blood transfusions (before 1985) and vaccine against hepatitis A and/or B) are a set of features that describe the data. The input data consists of values of the different attributes. Using these attribute values, the decision tree generates a class as the output for each input data. The top of the tree, or first node, is called the root node, intermediate nodes are the descendant or “hidden” layers and the last level of nodes are the leaf nodes, that contain the final classification.