A Brief History of Decision Tree Analysis
Decision Tree Analysis is a general, predictive modeling tool that has applications spanning a number of different areas (11). In general, decision trees are constructed via an algorithmic approach that identifies ways to split a data set based on different conditions (12). In a publication produced by SAS, the utility of decision tree analysis was described as follows: "Decision trees are a form of multiple variable (or multiple effect) analyses. All forms of multiple variable analyses allow us to predict, explain, describe, or classify an outcome (or target)...This multiple variable analysis capability of decision trees enables you to go beyond simple one-cause, one-effect relationships and to discover and describe things in the context of multiple influences. Multiple variable analysis is particularly important in current problem-solving because almost all critical outcomes that determine success are based on multiple factors" (12).
One model for performing decision tree analysis was created by J. Ross Quinlan at the University of Sydney and presented in his book Machine Learning, vol.1, no. 1, in 1975 (10). His first algorithm for decision tree creation was called the Iterative Dichotomiser 3 (ID3). This algorithm was created based on the principles of Occam's razor, with the idea of creating the smallest, most efficient decision tree possible (13). Quinlan went on to further develop this model with his creation of the C4.5 algorithm, and finally, the C5.0 algorithm (13).
One model for performing decision tree analysis was created by J. Ross Quinlan at the University of Sydney and presented in his book Machine Learning, vol.1, no. 1, in 1975 (10). His first algorithm for decision tree creation was called the Iterative Dichotomiser 3 (ID3). This algorithm was created based on the principles of Occam's razor, with the idea of creating the smallest, most efficient decision tree possible (13). Quinlan went on to further develop this model with his creation of the C4.5 algorithm, and finally, the C5.0 algorithm (13).
