Table III.
The Confusion Matrix describes the performance of our supervised machine learning classification model on a set 210 test images cropped from DP virtual slides for which the true HP diagnoses were known.
| Predicted HP negative | Predicted HP positive | Totals | |
|---|---|---|---|
| Actual HP negative | 123 | 18 | 141 |
| Actual HP positive | 7 | 62 | 69 |
| Totals | 130 | 80 | 210 |
Chi2(4, N = 210) = 116.74, p < 0.0001. Pearson’s Phi Coefficient = 0.75.