Home › Glossary › Data › Holdout data

Beginner · Data

Holdout data

Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.

TL;DR. Data intentionally set aside and not used during model training for evaluation.

Technical Definition

Data intentionally set aside and not used during model training for evaluation.

How it works

Holdout data refers to portions of a dataset that are deliberately excluded from the training process. These sections are typically split into validation and test sets. Using holdout data is crucial for assessing how well a model generalizes to new, unseen data, providing a more realistic estimate of its performance in real-world scenarios.

Related Concepts

  • Training Data — The portion of a dataset used to fit a model's parameters.
  • Generalization — A model's ability to perform well on new, unseen data — not just its training set.

Further Reading

  • Google ML Glossary