Our algorithm works for data that has a large amount of discrimination.

The feature data used does not have a large amount of discrimination.