Research Article
[Retracted] Arabic Speech Analysis for Classification and Prediction of Mental Illness due to Depression Using Deep Learning
Table 2
Performance comparisons of hybrid model with RNN and CNN.
| | | Accuracy (%) | AUC | Sensitivity (%) | Specificity (%) | FPR | FNR |
| Training | RNN | 80.70 | 0.81 | 100 | 61.9 | 0.0 | 0.380 | CNN | 88.50 | 0.89 | 95.6 | 816 | 0.043 | 0.183 | CNN + SVM | 90 | 0.90 | 98.5 | 81.6 | 0.014 | 0.183 |
| Testing | RNN | 81.60 | 0.81 | 100 | 62 | 0.0 | 0.379 | CNN | 86.60 | 0.86 | 93.5 | 79.3 | 0.064 | 0.206 | CNN + SVM | 91.60 | 0.91 | 100 | 82.7 | 0.0 | 0.172 |
|
|