This flashcard is just one of a free flashcard set. See all flashcards!

Cross validation is the idea to take the avaiable Data and split it into multiple parts and then use some of those parts as the training data and some as validation data. This is done over multiple rounds with the parts used for training and the parts used for validation changing.
This gives us insight on ho wthe model will generalize to an independent dataset in order to limit problems like overfitting.