Tag: how to handle duplicate data in machine learning