StackedML
Practice
Labs
Questions
Models
Pricing
Sign in
Questions
/
Model Evaluation & Experimentation
/
Validation
/
Cross-validation
← Previous
Next →
455.
Leave-One-Out Cross-Validation
medium
What is leave-one-out cross-validation (LOOCV)?
A
Cross-validation where the model is retrained from scratch on one random sample each iteration
B
A form of bootstrap validation that leaves one sample out of each bootstrap resample during each iteration of the procedure
C
K-fold cross-validation where k equals the number of training samples — each sample serves as its own validation fold
D
A special case of cross-validation where one class is held out as validation in each fold
Sign in to verify your answer
← Back to Questions