Back to Glossary

Machine Learning task

Classification

Assigning observations to one or more predefined classes using patterns learned from data.

Related domain: Data AnalyticsConcept ID: concept-classification

Definition

Classification is a Machine Learning task in which observations are assigned to one or more predefined classes using patterns learned from available data.

Human Explanation

A model estimates which predefined category or categories best match a new observation. It may also return a score or probability rather than a final operational decision.

Why it Matters

Classification supports scalable handling of tasks that require observations to be sorted into defined categories.

Conceptual Boundary

Classification predicts categorical outcomes rather than continuous numerical values. Binary, multiclass and multilabel Classification are related forms, and the predicted class does not by itself determine what action should be taken.

Practical Perspective

Applications include spam detection, image recognition, document classification and fraud detection. Class balance, error costs and the decision threshold can matter more than overall accuracy alone.

Related Concepts