Supervised, Unsupervised, and Reinforcement Learning — Machine Learning with Scikit-learn

Machine learning is the discipline of building systems that learn patterns from data rather than following explicitly programmed rules, and the first essential distinction is the type of learning signal the system receives. Supervised learning trains on labelled examples — input-output pairs — and learns to map new inputs to outputs; unsupervised learning finds structure in unlabelled data; and reinforcement learning learns by interacting with an environment and receiving rewards or penalties. These three paradigms exist because different problems offer different information: some domains have abundant labelled data, others have vast amounts of unlabelled data, and others offer only delayed feedback from actions. Understanding which paradigm fits a problem is the first decision a machine-learning practitioner makes, because it determines the algorithms available, the data required, and the evaluation approach — choosing the wrong paradigm leads to ill-defined problems and failed systems.

Analogy🏏Cricket

🏏 Think of it like cricket: A batting coach who teaches only one shot — the forward defensive — gives advice with high bias: it is consistently wrong for deliveries that demand a drive or a pull, regardless of how much practice the batsman does. A coach who memorises every ball of the batsman's training career gives advice with high variance: he predicts each training ball perfectly but fails completely on new balls from a different bowler, because he learned the noise of that specific bowler rather than the underlying principles. The great coach finds the balance — teaching the core principles that generalise, without over-specifying for the particular training environment. Just as great coaching lies between the extremes, great ML models balance bias and variance.