A Machine Learning Specialist was given a training dataset to create a classification model. The Specialist plans to partition this dataset into subsets of data which will be used to train five different versions of the model. The best model version shall be decided by its evaluation score. Which is the MOST effective approach for accomplishing this task?