In Amazon Machine Learning, what is the main function of data rearrangement? It allows you to include all the data when creating data source using the PercentEnd parameter. It enables you to create a datasource that is based on only a portion of the input data that it points to. It automatically reserves 70% of your data for ML model evaluation and uses the other 30% for training. It allows you to create complementary datasources for training and evaluation using PercentBegin parameter.