ristikvalidoinnin

Ristikvalidointi, or cross‑validation, is a statistical technique used to assess how a predictive model will generalise to an independent data set. The basic idea is to partition the available data into complementary subsets, train the model on one subset and validate it on the other. By repeating this process several times, one obtains an estimate of the model’s predictive performance that is less biased than a simple split of the data into a single training and test set.

The most common variant is k‑fold cross‑validation. The data set is divided into k roughly equal parts, or folds. For each of the k iterations, one fold is held out as the validation set while the remaining k‑1 folds are used to train the model. The performance metrics are then averaged over all iterations. A special case is leave‑one‑out cross‑validation (LOOCV), where k equals the number of observations; this provides an almost unbiased estimate but can be computationally expensive for large data sets.

Other variants include stratified k‑fold, which preserves the class distribution in classification problems, and repeated cross‑validation, where the k‑fold process is repeated multiple times with different random splits to reduce variance of the estimate.

Cross‑validation is widely used in machine learning, statistical modelling, and bioinformatics for model selection, hyper‑parameter tuning, and to detect overfitting. Its strength lies in providing a realistic estimate of out‑of‑sample error, though it requires more computational resources than a single test split. In practice, cross‑validation is implemented in popular libraries such as sklearn in Python, caret in R, and many others.