Table 4
Correlation removal pre-processing.
| Model | Accuracy | Precision | Recall | Deviation from dem. parity |
|---|---|---|---|---|
| Imp. feat. | 96.2% | 77.8% | 70.9% | 13.7% |
| Corr. Rem. | 96.2% | 77.4% | 70.9% | 0.07% |
Notes. Performance of the model trained on PFI-important features (top row) compared with an identical model trained on the same features preprocessed with CorrelationRemover (bottom row). The deviation from demographic parity was measured as the percent difference in the fraction of objects predicted as flares on the test set by either classifier.
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.