Table 1.
f1-scores on the test set of the ANN model trained on different mock datasets.
| Mock dataset | S/N | E(B − V) | Test f1-score |
|---|---|---|---|
| Infinite S/N | ∞ | 0 | 1.00 |
| High S/N | 100 | 0 | 0.97 |
| High S/N + extinction | 100 | [0 − 0.4] | 0.94 |
| Low S/N | 20 | 0 | 0.82 |
| Multi-label | 100 | [0 − 0.4] | 0.82 |
| Single spaxel | ⟨37⟩ ( * ) | ⟨0.16⟩ ( * ) | 0.80 |
| Integrated nebulae | 370 | ⟨0.15⟩ ( * ) | 0.89 |
Notes. The single spaxel and integrated nebulae entries are related to the models trained for the application to the M33 data. The f1-scores measured for both these two models and the multi-label one consider only the instances with a contribution from the prevalent class > 50%. The ‘( * )’ marks average values of the S/N or extinction distributions that have been tailored to match that of the real data, as described in Sect. 5.1.
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.