Chunk #24 — The Methods — How Do Classification and Regression Trees Work? — Splitting and Stopping

Source

An introduction to recursive partitioning: rationale, application, and characteristics of classification and regression trees, bagging, and random forests.

Embedded

yes

Text

Both the CART algorithm of Breiman et al. (1984) and the C4.5 algorithm (and its predecessor ID3) of Quinlan (1986, 1993) conduct binary splits in numeric predictor variables, as depicted in Figure 1. In categorical predictor variables (of nominal or ordinal scale of measurement) C4.5 produces as many nodes as there are categories (often referred to as “k-ary” or “multiple” splitting), while CART again creates binary splits between the ordered or unordered categories. We concentrate on binary splitting trees in the following and refer to Quinlan (1993) for k-ary splitting.