Decision trees are a popular algorithm that allows to identify simple rules that allow to identify the class of items in a dataset. The decision paths that are formed by the sequential application of these rules, form a so-called decision tree, that unmixes the element that are in the dataset.
We show here decision trees that were generated using the python library scikit-learn and the visualization framework d3.js. On toop of the tree, in the root node, the items from all classes in the dataset are mixed. As more rules are applied, and the tree branches out to the bottom, the classes are better and better unmixed.
Hovering with the mouse over the paths allows to see the rules that were applied to unmix certain classes.