Confusion Matrix
- Measures the test performance of a classification system on a per-class basis by indicating the number of samples of actual class a predicted as class b.
- The rows relate to the actual class labels a and the columns to the predicted class labels b.