What is Classification?
Classification is a supervised machine learning task where a model evaluates an input and predicts which specific, predefined category or label it belongs to.
How does it work?
Classification models are trained on datasets that already have correct labels attached. Types of classification include:
- Binary classification: Choosing between exactly two options (e.g., classifying an email as Spam or Not Spam, or a credit card transaction as Fraudulent or Legitimate).
- Multiclass classification: Choosing exactly one category from a larger list (e.g., determining if an image shows a dog, a cat, or a bird).
- Multilabel classification: Assigning multiple categories to a single input (e.g., tagging a news article as both "Politics" and "Economics").
What is it commonly confused with?
Distinguish classification from clustering. Classification predicts known labels based on past examples. Clustering attempts to discover unknown groups in unlabeled data.
Why does it matter?
Classification is one of the most commercially valuable applications of AI. It automates massive volumes of routing, moderation, and diagnostic tasks across industries, from identifying toxic comments on social media to flagging tumors in medical scans.