Table 1.
Average Performance Results Across 10 Repeats
| Classification with Features Modeled Individually | |||||||
|---|---|---|---|---|---|---|---|
| Model | Acc | bAcc* | Sens* | Spec* | PPV* | F1 | Gmean |
| PTSD vs. Controls | 68% | 67% | 56% | 77% | 65% | 0.60 | 66% |
| ELS+ vs. ELS−, PTSD | 64% | 63% | 56% | 70% | 58% | 0.54 | 63% |
| ELS+ vs. ELS−, Controls | 68% | 57% | 30% | 84% | 42% | 0.33 | 50% |
| Classification with EPIC | |||||||
| Model | Acc | bAcc* | Sens* | Spec* | PPV* | F1 | Gmean |
| PTSD vs. Controls | 71% | 69% | 58% | 81% | 69% | 0.62 | 69% |
| ELS+ vs. ELS−, PTSD | 66% | 64% | 53% | 75% | 60% | 0.54 | 63% |
| ELS+ vs. ELS−, Controls | 70% | 62% | 43% | 81% | 48% | 0.44 | 59% |
Abbreviations and formulas: Acc = Accuracy, true positive (TP)+true negative (TN)/TP+TN+false positive (FP)+false negative (FN); bAcc = balanced accuracy, [(TP/(TP+FN)+TN/(TN+FP)]/2; F1 = harmonic mean of precision and recall, 2*(Recall * Precision)/(Recall + Precision); Gmean = geometric mean, √TPrate*TNrate; *reported in main text; PPV = positive predictive value (precision), TP/TP+FP; Sens = sensitivity (recall), TP rate, TP/TP+FN; Spec = specificity, TN rate, TN/FP +TN.