How do supervised, unsupervised, and reinforcement learning differ?

Supervised learning learns from examples paired with target labels, such as emails labeled spam or legitimate. Unsupervised learning searches for structure without those target labels, such as customer clusters. Reinforcement learning learns a policy through actions and rewards over time, often when an action affects future opportunities. Choose according to available feedback and the decision being made. A clustering result is not automatically a meaningful business segment, and reinforcement learning is usually unnecessary when a simpler supervised prediction directly solves the problem.