Chapter 10CART: Classifi cation and Regression TreesDan SteinbergContents10.1Antecedents ......................................................... 18010.2Overview ........................................................... 18110.3A RunningExample................................................. 18110.4TheAlgorithmBriefl yStated........................................ 18310.5SplittingRules ...................................................... 18510.6PriorProbabilitiesandClassBalancing............................... 18710.7MissingValue Handling............................................. 18910.8Attribute Importance ................................................ 19010.9DynamicFeatureConstruction....................................... 19110.10 Cost-Sensitive Learning ............................................. 19210.11 StoppingRules,Pruning,Tree Sequences, and Tree Selection ......... 19310.12 ProbabilityTrees .................................................... 19410.13 TheoreticalFoundations ............................................. 19610.14 Post-CART Related Research ........................................ 19610.15 Software Availability ................................................ 19810.16 Exercises ............................................................ 198References.................................................................. 199The1984monograph,“CART: Classification and Regression Trees,” coauthored byLeo Breiman, Jerome Friedman, Richard Olshen, and Charles Stone (BFOS), repre-sents a major milestone in the evolution of artificial intelligence, machine learning,nonparametric statistics, and data mining. The work is important for the compre-hensiveness of its study of decision trees, the technical innovations it introduces, itssophisticated examples of tree-structured data analysis, and its authoritative treatmentof large...
