# How does the ID3 Algorithm differ from a generic Decision Tree learning algorithm

Based on the notes of my Machine Learning lecturer, I am struggling to understand how the two algorithms differ? Both seem to select the most informative feature A (based on least entropy), then create subsets of the example set for each value of A. Then create a new branch on the tree by recursively calling ID3 on each subset. Could someone please explain how they differ?

• After thinking more about this, I'm beginning to think maybe that the decision tree itself is an abstract concept, and that ID3 is just an extension/implementation of a "decision tree". Is this correct? Jan 6, 2015 at 16:37

## 1 Answer

The ID3 algorithm is an algorithm for inferring a decision tree: given a training set, it tries to find a decision tree that tends to label a large fraction of the training set accurately. See, e.g., https://en.wikipedia.org/wiki/ID3_algorithm.

There is no such thing as "the generic decision tree learning algorithm". There is no such generic algorithm.

There are many algorithms for learning a decision tree. ID3 is probably the most well-known algorithm for decision tree learning, and the simplest, but there are other algorithms. Other algorithms include C4.5, CART, and others.