Decision Tree Learning (ID3 / C4.5)

Why It Was Important

Decision trees were the first widely used machine-learning method whose output a human could read and check, which made them the practical bridge between expert systems and statistical learning. Quinlan's ID3 and its successor C4.5 became among the most-used algorithms in data mining, and the technique underpins the random forests and gradient-boosted ensembles that still win on tabular data.

Who Invented It

Ross Quinlan, Leo Breiman

Quinlan built ID3 and C4.5; Breiman and colleagues developed the CART framework in parallel.

Applications

• Credit Scoring
• Medical Diagnosis
• Data Mining
• Ensemble Methods