Types of Machine Learning

Supervised, unsupervised, and reinforcement learning — explained simply.

Supervised Learning

The model learns from labeled data — input-output pairs — to predict outputs for new inputs. Examples: spam email classification, house price prediction.

Unsupervised Learning

The model finds patterns in unlabeled data without predefined outputs. Examples: customer segmentation (clustering), and dimensionality reduction for visualization.

Reinforcement Learning

An agent learns by interacting with an environment, receiving rewards or penalties for its actions, and gradually improving its strategy — used in game-playing AI and robotics.

🌍 Real-World Use

Netflix/YouTube recommendations use a mix: supervised learning predicts what you'll rate highly based on past behavior, while unsupervised learning groups similar users/content together for better suggestions.

💡 Pro Tip

The fastest way to distinguish these in an interview: 'Supervised = has labeled answers, Unsupervised = finds patterns with no answers given, Reinforcement = learns from rewards/penalties through trial and error.'

🧪 Quick Self-Test

Check what you just learned — no pressure, just practice.

1. Which type of ML uses labeled data?

2. What guides a Reinforcement Learning agent's improvement?

← Back to all Notes