Table 1
Data coverage in spectroscopic redshift and sample division.
| ɀmin | ɀmax | Number of ɀ bins | Bin size | Training sample | Validation sample | Target sample | |
|---|---|---|---|---|---|---|---|
| SDSS | 0.0 | 0.4 | 180 | 2.22 × 10–3 | 393 219 | 20 000 | 103 305 (*) |
| CFHTLS | 0.0 | 4.0 | 1000 | 4.0 × 10–3 | 100 000+23 434 (**) | 14 759 | 20 000 |
| KiDS | 0.0 | 3.0 | 800 | 3.75 × 10–3 | 100 000 | 14 147 | 20 000 |
Notes. (*)50% of the galaxies from the SDSS target sample are used as a reference subsample for the refitting procedure; the remaining 50% are held out for inference (Sect. 3.5). (**)The 23 434 galaxies with low-resolution spectroscopic redshifts from the CFHTLS training sample are used in supervised contrastive learning (SCL), but not in the downstream procedures.
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.