Table 13 Unimodal Text results based on Recursive Feature Elimination (RFE) with the Logistic Regression model and with re-scaling
From: Multimodal machine learning for language and speech markers identification in mental health
Model | Features | Accuracy | AUC-ROC | F1 - 0s | F1 - 1s |
|---|---|---|---|---|---|
SVM | 10 | 80.40% | 87.09% | 0.86 | 0.67 |
SVM | 15 | 81.97% | 89.38% | 0.87 | 0.70 |
SVM | 20 | 86.77% | 93.33% | 0.91 | 0.78 |
SVM | 25 | 84.68% | 92.85% | 0.89 | 0.76 |
RF | 10 | 78.27% | 85.57% | 0.85 | 0.60 |
RF | 15 | 83.60% | 85.16% | 0.89 | 0.69 |
RF | 20 | 78.28% | 82.09% | 0.86 | 0.57 |
RF | 25 | 85.72% | 91.75% | 0.90 | 0.73 |
LogReg | 10 | 80.92% | 87.94% | 0.86 | 0.68 |
LogReg | 15 | 83.57% | 89.52% | 0.88 | 0.73 |
LogReg | 20 | 87.82% | 92.44% | 0.91 | 0.79 |
LogReg | 25 | 85.72% | 93.95% | 0.90 | 0.76 |
Dense Layers | 10 | 84.65% | 88.94% | 0.88 | 0.73 |
Dense Layers | 15 | 82.05% | 87.81% | 0.88 | 0.72 |
Dense Layers | 20 | 83.57% | 88.01% | 0.87 | 0.67 |
Dense Layers | 25 | 84.11% | 91.79% | 0.89 | 0.74 |