Confusion matrix for the ‘1 vs all’ classification technique, when applied to the small well-represented dataset
| . | Predicted class . | ||||||
|---|---|---|---|---|---|---|---|
| Infiltration . | Joint . | Deposits . | Roots . | Crack . | Surface . | ||
| Correct class | Infiltration | 0.5 | 29.7 | 0.9 | 0.9 | 1.8 | 0.9 |
| Joint | 0.7 | 89.7 | 4.4 | 0.4 | 2.9 | 1.8 | |
| Deposits | 0.8 | 21.3 | 71.3 | 0.8 | 0.8 | 4.8 | |
| Roots | 1.1 | 24.2 | 9.1 | 60.2 | 2.2 | 3.2 | |
| Crack | 1.3 | 27.9 | 3.5 | 0.9 | 62.8 | 3.5 | |
| Surface | 1.3 | 28.0 | 7.1 | 1.8 | 4.0 | 57.8 | |
| . | Predicted class . | ||||||
|---|---|---|---|---|---|---|---|
| Infiltration . | Joint . | Deposits . | Roots . | Crack . | Surface . | ||
| Correct class | Infiltration | 0.5 | 29.7 | 0.9 | 0.9 | 1.8 | 0.9 |
| Joint | 0.7 | 89.7 | 4.4 | 0.4 | 2.9 | 1.8 | |
| Deposits | 0.8 | 21.3 | 71.3 | 0.8 | 0.8 | 4.8 | |
| Roots | 1.1 | 24.2 | 9.1 | 60.2 | 2.2 | 3.2 | |
| Crack | 1.3 | 27.9 | 3.5 | 0.9 | 62.8 | 3.5 | |
| Surface | 1.3 | 28.0 | 7.1 | 1.8 | 4.0 | 57.8 | |
Bold values identify the True Positive Rates for each fault type.