Unbalanced sample causes issues and more efforts as we know.
When I am handling the issue， I am confused about the definition. Say, I have a training dataset of 200
