Chunk #42 — The Methods — How Do Classification and Regression Trees Work? — Prediction and Interpretation of Classification and Regression Trees

Source

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

Embedded

yes

Text

We will further illustrate this important issue and source of misinterpretations by means of varying the effects in our artificial data set. The resulting classification trees are given in Figure 4. Only the left plot in Figure 4, where the effect of alcohol_per_month is the same in both branches created by friends_smoke, represents two main effects of alcohol_per_month and friends_smoke without an interaction: The main effect of friends_smoke shows in the higher relative frequencies of “yes” answers in nodes 6 and 7 as compared to nodes 3 and 4. The main effect of alcohol_per_month shows in the higher relative frequencies of “yes” answers in nodes 4 and 7 as compared to nodes 3 and 6 respectively.