Table 2 Performance of Model predicting binary shock index using manually calculated CPD.(Coeff—Model coefficient of difference percent, AUC—Area Under the curve, PPV—Positive predictive value, NPV—Negative predictive value, Cut-off—Average Threshold of Ten Folds, NS- Non-shock, S- Shock).
From: Predicting Hemodynamic Shock from Thermal Images using Machine Learning
Time-point, (NS, S) | Coeff | AUC | Accuracy | Sensitivity | Specificity | PPV | NPV | Cut-off |
|---|---|---|---|---|---|---|---|---|
Mean(SE) | Mean(SE) | Mean(SE) | Mean(SE) | Mean(SE) | Mean(SE) | Mean | ||
0 hr (detection), (146, 107) | 0.03 | 0.79 (0.02) | 0.75(0.02) | 0.69(0.05) | 0.79(0.05) | 0.72(0.05) | 0.8(0.03) | 0.47 |
3 hr (prediction), (141, 107) | 0.01 | 0.79(0.04) | 0.74(0.04) | 0.72(0.05) | 0.78(0.07) | 0.73(0.06) | 0.81(0.03)) | 0.45 |
6 hr (prediction), ((129, 109)) | 0.01 | 0.65(0.04) | 0.66(0.02) | 0.48(0.09) | 0.81(0.06) | 0.72(0.07) | 0.71(0.05) | 0.62 |
12 hr (prediction), (131, 118) | 0.01 | 0.7(0.03) | 0.69(0.02) | 0.67(0.03) | 0.68(0.04) | 0.65(0.06) | 0.7(0.03) | 0.41 |
