Classification
Last Updated: July 29, 2026 | By Mihail Sebastian | AI Dictionary
A supervised learning task that assigns inputs to predefined categories, such as spam or not spam, based on patterns learned from labeled examples.
What is Classification?
Classification is a supervised learning task in which a model learns from labeled examples to assign new inputs to predefined categories, such as spam or not spam. The categories are fixed in advance, and every training example carries the correct answer.
The word also has non-ML senses: the EU AI Act sorts AI systems into risk classes, and documents carry security classifications. Those are human-made sorting schemes, not the machine learning technique this page covers.
Types of Classification
- Binary classification chooses between two classes: an email is spam or it isn’t, a transaction is fraudulent or legitimate.
- Multiclass classification picks one class from several, such as sorting news articles into sports, politics, or technology.
- Multilabel classification lets one input carry several labels at once, like a photo tagged “beach,” “sunset,” and “vacation” together.
Classification vs Regression Analysis
Classification predicts which category an input belongs to; regression analysis predicts a number. If the question is “which kind?”, you want a classifier; if it is “how much?”, you want regression. Both learn from labeled data, and the same dataset supports either, depending on what you ask of it.
Clustering looks similar but is a different job: it finds groups in unlabeled data, with no answer key to learn from.
| Classification | Regression Analysis | |
|---|---|---|
| Predicts | A category label | A continuous number |
| Typical question | “Is this transaction fraud?” | “What will this house sell for?” |
| Output | A class, usually with a confidence score | A numeric value |
| Evaluated with | Accuracy, precision, recall | Error measures such as MSE or MAE |
Example of Classification
A spam filter is classification end to end. The model trains on millions of emails labeled spam or legitimate, and learns which signals separate the two: sender reputation, suspicious links, wording patterns. When a new email arrives, the model scores it and routes anything above a threshold to the spam folder.
The two error types cost different amounts. Letting spam through is an annoyance; burying a real job offer in the spam folder is a genuine loss. That is why classifiers are judged on precision and recall separately, not on accuracy alone.
Related AI terms: Regression Analysis · Clustering · Sentiment Analysis · Logistic Regression · Confusion Matrix
Did you like the Classification gist?
Learn about 250+ need-to-know artificial intelligence terms in the AI Dictionary.
Mihail Sebastian — Writes about AI governance, regulation, and the technology behind them. Placeholder bio — replace with a real credential line. About