Cross-Validation
A model evaluation technique that splits data into multiple folds, trains on some folds and tests on the held-out fold, repeating so every fold serves as the test set. It provides a robust estimate of model performance.
Why It Matters
Cross-validation gives you confidence in your model's performance. A model that scores well on one test split might have gotten lucky — cross-validation averages over multiple splits.
Example
5-fold cross-validation: split 1,000 examples into 5 groups of 200, train on 4 groups and test on 1, repeat 5 times so each group gets to be the test set.
Think of it like...
Like a restaurant that tests a new dish with five different groups of diners rather than just one — the average feedback is much more reliable.
Related Terms
Validation Data
A subset of data used during training to tune hyperparameters and monitor model performance without touching the test set. It acts as an intermediate checkpoint between training and final evaluation.
Test Data
A separate portion of data held back from training that is used to evaluate a model's performance on unseen examples. Test data provides an unbiased estimate of how well the model will perform in the real world.
Overfitting
When a model learns the training data too well — including its noise and random fluctuations — and performs poorly on new, unseen data. The model essentially memorizes rather than generalizes.
Hyperparameter Tuning
The process of systematically searching for the best combination of hyperparameters for a model. Since hyperparameters are set before training, finding optimal values requires experimentation.