Holdout Dataset

A Holdout Dataset is a learning evaluation dataset that is not in the training task during an in-sample model evaluation task (by a learning system).

• Example(s):
  • a randomly selected 10% sample from the Iris dataset.
  • …
• Counter-Example(s)
  • Training Dataset,
  • Test Dataset.
• See: Holdout Evaluation, Sample Dataset, Cross-Validation Task, Model Evaluation Task.

---

---

References

2018

• (MLG, 2018) ⇒ holdout data. In: Machine Learning Glossary: https://developers.google.com/machine-learning/glossary/ Retrieved: 2018-05-20
  • QUOTE: Examples intentionally not used ("held out") during training. The validation data set and test data set are examples of holdout data. Holdout data helps evaluate your model's ability to generalize to data other than the data it was trained on. The loss on the holdout set provides a better estimate of the loss on an unseen data set than does the loss on the training set.

2017

• (Sammut & Webb, 2017) ⇒ (2017) "Holdout Set". In: Sammut C., Webb G.I. (eds) Encyclopedia of Machine Learning and Data Mining. Springer, Boston, MA
  • QUOTE: A holdout set is a data set containing data that are not used for learning and that are used for evaluation by a learning system.

---