Decision trees

Classification task using decision trees.

---

WHAT
Classification of points with two features: (i) diameter size and (ii) colour.

HOW
The decision tree looks at all features in turn and splits the data using the one which bring about the larger amount of information to help with the classification task.

The algorithm stops when:

• all data points have been correctly classified;

• there are no (more) attributes to split upon;

• the information gain originating from the split is very limited; or

• the tree reaches a maximum (pre-defined) depth.

OUTPUT
1 tree.
2 splits: a. diameter size; b. colour.
And all data points have been correctly classified (: