Fundamentals

Key Terms

• Feature: an input column (age, clicks, price).
• Label/Target: what you predict (class or number).
• Model: a function with learned parameters.
• Loss: how wrong predictions are.
• Generalization: performance on unseen data.

# Typical flow
# data -> features -> model.fit(train) -> evaluate(valid/test) -> deploy -> monitor

Supervised vs Unsupervised

• Supervised: you have labels (classification/regression).
• Unsupervised: no labels (clustering, dimensionality reduction).