Table 2.
Model performance.
| Set of features | Dimensionality reduction | Predictor | RMSE (std) |
|---|---|---|---|
| Progressing features (29 features) | Feature selection (2 features) | LR (model1) | 11.86 (0.67) |
| Feature selection (6 features) | LR (model2) | 11.17 (0.80) | |
| All progressing (29 features) | RF (model3) | 10.02 (0.88) | |
| PCA (10 factors) | LR (model4) | 11.25 (0.68) | |
| RF (model5) | 10.92 (0.65) | ||
| Original features (122 features) | PCA (31 factors) | LR (model6) | 10.80 (0.91) |
| RF (model7) | 10.32 (0.76) |
The table shows the regression results when automatically selecting features.
RMSE Root Mean Square Error, std standard deviation.