Labeled Data

In labeled datasets, each data sample is paired with a known label that identifies its correct outcome. For example, an email may be labeled as "spam" or "not spam," an image may be labeled as "cat" or "dog," or a medical scan may include annotations indicating the presence of a condition. These labels are typically created by human annotators, domain experts, or trusted automated processes.

Supervised machine learning models rely on labeled data to learn how to make accurate predictions. The quality and consistency of the labels directly influence model performance. Well-labeled datasets help improve accuracy, reduce training errors, and provide a reliable foundation for evaluating model performance.