Compare the individual models with each other to see how they performed in the test for a specific use case.