Classification is a supervised learning task where a model predicts a discrete label for each input. Common examples include spam detection (spam vs. not spam) and image classification (various object categories). Models can be evaluated with confusion matrices, accuracy, precision, and recall.