Train/validation/test sets

In machine learning, these are three distinct subsets of a dataset used during the model development process. The training set is used to train the model, allowing it to learn from the data. The validation set is used to tune the model's parameters and prevent overfitting. The test set is used after the model has been finalized to assess its performance on new, unseen data.