Table 3.
The results of five models based on optimal internal personal and external features for depressive symptoms on the training and test sets.
| Data sets | Model | Accuracy | Precision | Recall | F1-score | AUC |
|---|---|---|---|---|---|---|
| Training set | KNN | 0.785 ± 0.03 | 0.779 ± 0.03 | 0.785 ± 0.03 | 0.777 ± 0.03 | 0.851 ± 0.03 |
| GNB | 0.795 ± 0.02 | 0.802 ± 0.01 | 0.795 ± 0.02 | 0.797 ± 0.02 | 0.871 ± 0.02 | |
| SVM | 0.803 ± 0.02 | 0.799 ± 0.02 | 0.803 ± 0.02 | 0.792 ± 0.02 | 0.874 ± 0.02 | |
| RF | 0.787 ± 0.01 | 0.781 ± 0.01 | 0.787 ± 0.01 | 0.780 ± 0.01 | 0.870 ± 0.02 | |
| LR | 0.798 ± 0.01 | 0.794 ± 0.01 | 0.798 ± 0.01 | 0.790 ± 0.01 | 0.875 ± 0.02 | |
|
| ||||||
| Test set | KNN | 0.798 | 0.791 | 0.798 | 0.788 | 0.871 |
| GNB | 0.791 | 0.803 | 0.791 | 0.794 | 0.893 | |
| SVM | 0.808 | 0.803 | 0.808 | 0.799 | 0.894 | |
| RF | 0.815 | 0.810 | 0.815 | 0.810 | 0.892 | |
| LR | 0.801 | 0.795 | 0.801 | 0.803 | 0.895 | |