Table 2 Performance metrics of each random forest model by response type comparison. PMC is the probability of misclassification by the model, and accuracy is 1– PMC.
Response Type Model | Number of each Response Type (n: n) | Sensitivity (%) | Specificity (%) | PMC (%) | Accuracy (%) |
|---|---|---|---|---|---|
ANR vs. FP | 19,639: 5,553 | 61.4 | 63.9 | 37.0 | 63.0 |
NRQ vs. FP | 14,043: 5,553 | 58.9 | 91.5 | 18.0 | 82.0 |
NRQ vs. FP (complete case) | 9,451: 5,546 | 59.4 | 86.6 | 23.3 | 76.7 |
QO vs. FP | 647: 5,553 | 99.4 | 10.7 | 9.8 | 90.2 |