Table 7 Performance metrics of four classification models (Discriminant, Naive bayes, Multi-Class SVM, and DNN) in the binary classification of anxiety, evaluated based on recall (TPR), false negative rate (FNR), precision (PPV), false discovery rate (FDR), F1-score, and overall accuracy for both anxious and no anxiety classes.
| Â | Â | TPR (Recall) % | FNR (false neg rate) % | PPV (precision) % | FDR (false disc rate) % | F1 score % | Overall accuracy % |
|---|---|---|---|---|---|---|---|
LDA | Anxious | 62.59 | 37.40 | 61.45 | 38.54 | 62.01 | 61.67 |
No anxiety | 60.74 | 39.25 | 61.88 | 38.11 | 61.30 | ||
Naive Bayes | Anxious | 55.92 | 44.07 | 59.44 | 40.55 | 57.63 | 58.89 |
No anxiety | 61.85 | 38.14 | 58.39 | 41.60 | 60.07 | ||
Multi-class SVM | Anxious | 61.85 | 38.14 | 58.80 | 41.19 | 60.28 | 59.26 |
No anxiety | 56.66 | 43.33 | 59.76 | 40.23 | 58.17 | ||
DNN | Anxious | 58.88 | 41.11 | 58.02 | 41.97 | 58.45 | 58.15 |
No anxiety | 57.40 | 42.59 | 58.27 | 41.72 | 57.83 |
