Table 2
Summary of test stage for the different models.
| F.Extraction | ML-Model | R2 | MAE | RMSE |
|---|---|---|---|---|
| – | SVM | 0.479 ± 0.007 | 0.182 ± 0.002 | 0.214 ± 0.002 |
| Standard scaling | Random Forest | 0.942 ± 0.002 | 0.029 ± 0.001 | 0.070 ± 0.001 |
| t-test feature selection | XGBoost | 0.930 ± 0.001 | 0.047 ± 0.007 | 0.078 ± 0.001 |
| t-test feature selection | LightGBM | 0.928 ± 0.002 | 0.039 ± 0.001 | 0.079 ± 0.002 |
| – | k-NN | 0.714 ± 0.004 | 0.102 ± 0.001 | 0.158 ± 0.001 |
Notes. In addition to RMSE we include additional metrics to evaluate the results, the coefficient of determination (R2-score) and Mean Absolute Error (MAE). Both RMSE and MAE are minimized to 0, while the R2-score is maximized to 1 (i.e. the predictions are perfect). Note that the best ML method is bold faced.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.