You are solving a classification task and need to evaluate your model on a limited data sample using k-fold cross-validation. You start by configuring the k parameter as the number of splits.

Which value should you use for the k parameter?