Beta. Content is under active construction and has not been peer-reviewed. Report errors on
GitHub
.
Disclaimer
Theorem
Path
Curriculum
Paths
Demos
Diagnostic
Search
Quiz Hub
/
Model Evaluation Best Practices
Model Evaluation Best Practices
3 questions
Difficulty 2-4
View topic
Foundation
0 / 3
2 foundation
1 intermediate
Adapts to your performance
1 / 3
foundation (2/10)
conceptual
Best practice is to split data into train, validation, and test sets. Why three splits rather than just train and test?
Hide and think first
A.
Validation provides cross-validation folds, while test provides a separate dataset that is held out during training only
B.
Validation is used for hyperparameter selection; test is held out for final evaluation to avoid overfitting to the evaluation set itself
C.
Two splits are sufficient and historically standard; the three-way split is a modern marketing practice
D.
Training and test split is for shallow models; validation is added specifically for neural networks that need extra evaluation
Submit Answer