Random Trees Overview

Random Trees is an ensemble model consisting of multiple C&R Tree-like trees. Each tree is grown on a bootstrap sample which is obtained by sampling the original data instances with replacement. Moreover, during tree growth, for each node the best split feature or predictor is selected from a specified smaller number of features which are drawn randomly from the full set. Each tree is grown to the largest extent possible, with no pruning. In scoring, Random Trees combines individual tree scores by majority voting (for classification) or average (for regression).

Model accuracy estimates are calculated in two ways in Random Trees. During model estimation, so-called “out-of-bag” instances (those not included in a particular bootstrap sample) are used to produce error estimates that are likely to better approximate errors found in predicting new data instances. These estimates are displayed in the Model Information table. Accuracy measures from the completed ensemble model on the full set of data are produced in the Model Evaluation panel using calculations that can be compared with those from other nodes shown in Model Evaluation panels.

Next steps

Like your visualization? Why not deploy it? For more information, see Deploy a model.