Reference · Glossary

Classification

Predicting a **discrete label** — spam/not, yes/no, cat/dog — from inputs or scores.

When to use

Routing tickets, filtering content, or any “pick a bucket” decision in ML intro labs.

When not to

When the target is a continuous number — that’s regression (price, temperature).

Example

Scores ≥ 0.5 → label 1 (approve), else 0 — a binary classifier with one threshold.