Table 7
Details of the LSST-like systems recovered by models trained on different controlled sets.
| Model trained on | ||||
|---|---|---|---|---|
| set A | set B | set C | ||
| Recovery | all | 28.0% | 28.1% | 30.2% |
| double | 29.5% | 29.2% | 31.3% | |
| quad | 21.9% | 23.3% | 25.9% | |
| ∆tmed of recovered systems | all | 4.5 | 4.4 | 4.1 |
| double | 4.8 | 4.7 | 4.6 | |
| quad | 2.6 | 2.1 | 1.6 | |
Notes. The top part presents the percentage of mock LSST unresolved systems recovered by models trained on different sets – A, B, and C. The three subrows show the recovery rate for the entire recovered sample, and then divided into doubles and quads. The bottom part shows the median time delays of the recovered systems, again divided into three sets – all, doubles, and quads from top to bottom. For comparison, the input LSST sample has ∆tmed = 2.0, 2.4, 1.1 days, respectively, for the full sample, doubles, and quads. It is evident that the models can recover the doubles better than the quads since the former systems have typically larger time delays.
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.