Recursive Partitioning

Recursive partitioning (RP), or Classification and Regression Tree (CART) analysis, represents an exploratory statistical method for the multivariable data-derived and computer-mediated analysis of uncovering a structure in a data set. It produces a graphical output in the form of a decision tree, which facilitates data interpretation. This graphical output may be used in medical decision making to stratify patients into risk categories. Unlike other individual modeling techniques such as nomograms or artificial neural networks, CART predicts a risk group or stratum on an individual basis.

Tree-structured methodologies were first put forward by Morgan and Sonquist in 1963 and Morgan and Messenger in 1973. The CART program for classification and regression trees implements an expanded and strengthened tree-structured approach due to Breiman and colleagues in 1984. Basic ...

  • Loading...
locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles