What is a 'Weak Learner'?

What is a 'Weak Learner'?

Approach: Define a model that is only slightly better than random guessing.

A learning algorithm that is only slightly better than random guessing (e.g., a decision tree with only one split, called a stump). Boosting algorithms combine these into a strong learner.

Follow-up: Can a weak learner have 100% accuracy on a subset?

Key concepts: stumps, boosting, accuracy.