# Cross-validation and stratified k-fold

Source: https://onco.cc/terms/cross-validation/  
OnCo record `cross-validation` (Term). Data CC BY-NC 4.0, attribute "Data from OnCo (onco.cc)"; commercial use needs a licence.

## TL;DR

Cross-validation splits the data into k folds, trains on k minus one and tests on the last, rotating so every sample is tested once; stratified folds keep the class balance the same in each fold.

## Summary

Cross-validation is a family of model validation techniques for assessing how a statistical analysis will generalise to independent data, using resampling and sample splitting (Wikipedia); stratifying the folds (Wikipedia on stratified sampling) keeps rare classes represented. It estimates performance but is not a substitute for an external cohort, and every preprocessing step (scaling, PCA, feature selection) must be fitted inside the training fold or the estimate leaks. Reporting the seed makes the folds reproducible.

## Fields

- Kind: Term
- Last checked: 2026-09-24
- Also known as: cross-validation; cross validation; k-fold cross-validation; 5-fold cross-validation; stratified k-fold; stratified cross-validation; nested cross-validation; random seed
- Tags: cansim-terms

## Notes

- Listed in the CanSim terms map 1.0.0 (docs/onco/terms.json, generated 2026-09-24), CC BY 4.0, attribution: CanSim project, an open, public-data-first cancer foundation-model programme; CanSim page path /terms/cross-validation.

## Sources

- Wikipedia: https://en.wikipedia.org/wiki/Cross-validation_(statistics)
- Wikipedia: stratified sampling: https://en.wikipedia.org/wiki/Stratified_sampling
- Wikipedia: https://en.wikipedia.org/wiki/Cross-validation_(statistics)

## Connected records

- terms: [Bootstrap resampling](https://onco.cc/terms/bootstrap/), [Cancer AI vocabulary (CanSim terms map)](https://onco.cc/terms/cancer-ai-vocabulary/), [Data leakage in model evaluation](https://onco.cc/terms/data-leakage/), [External validation](https://onco.cc/terms/external-validation/), [Train, validation and test split discipline](https://onco.cc/terms/train-test-discipline/)

---
JSON: https://onco.cc/api/v1/entities/cross-validation.json