Example

In a Regression Tree, each leaf represents a numeric value

Classification Trees have True or False or some other discrete category in each leaf.

Data

Non linear data:

Since each leaf corresponds to the average Drug Effectiveness in a different cluster of observations, the tree does a better job reflecting the data than the straight line:

Build Regression Tree from top to down

We start with a simple dataset (one variable):

Why did we start the tree with Dosage in the Root Node?

We split data at the first data point:

For each side of the split we calculate the average Drug Effectiveness. The values in each leaf are the predictions that this simple tree will make for Drug Effectiveness.