Figure 1: Bagging applied to regression using a regression tree. | Nature Methods

Figure 1: Bagging applied to regression using a regression tree.

From: Ensemble methods: bagging and random forests

Figure 1

(a) A sample of size n = 30 generated from a parabola (dashed line) with added noise and the associated fit from a regression tree (solid black line). (b) Four different bootstrap samples from the sample in a and their corresponding regression tree predictions. Solid points are in the bootstrap sample, and some are represented more than once. Hollow points are not in the bootstrap sample and are called out-of-bag (OOB) points. (c) Ensemble regression (blue line) formed by averaging bootstrap regressions in b. The original regression tree fit from a is also shown.

Back to article page