Open Access
Table 2
Comparison between VAE and DR3 classifications.
| Class | VAE all inputs | DR3 | # Sources |
|---|---|---|---|
| δ Cep |t2 Cep | 81.0(1.3) | 98.1 | 106 |
| RR Lyr | 92.9(0.4) | 96.7 | 259 |
| ECL | 92.3(0.2) | 94.8 | 651 |
| LPV | 97.7(0.2) | 98.9 | 1329 |
| δ Sct |γ Dor | 89.7(0.4) | 93.5 | 383 |
| RS | 0.0(0.0) | 0.0 | 183 |
| SOLAR | 88.4(0.4) | 75.1 | 262 |
| ACV|CP | 87.3(0.7) | 55.0 | 12 |
| YSO | 52.3(1.4) | 90.4 | 37 |
| AGN | 97.2(1.1) | 100.0 | 74 |
| Macro avg. | 81.5(1.1) | 80.3 | 3300 |
| Weighted avg. | 88.4(0.3) | 89.5 | 3300 |
Notes. F1−scores per class for the LR trained using all the latent variables (with merged subclasses) and the DR3 predictions from Rimoldini et al. (2023) after removing the sources belonging to the training set of the latter.
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.