Table 3.
Median continuous ranked probability score (CRPS) for the different algorithms using the Euclid selection for the weighted spectroscopic and L15 samples.
| Spec. sample | L15 sample | |
|---|---|---|
| Le Phare | 0.057 | 0.056 |
| CPz | 0.087 | 0.091 |
| Phosphoros | 0.082 | 0.083 |
| EAzY | 0.050 | 0.048 |
| METAPHOR | 0.036 | 0.034 |
| ANNz | 0.115 | 0.124 |
| GPz | 0.116 | 0.113 |
| GBRT | 0.166 | 0.165 |
| RF | 0.107 | 0.119 |
| Adaboost | 0.078 | 0.090 |
| DNF | 0.076 | 0.072 |
| frankenz | 0.071 | 0.072 |
| NNPZ | 0.084 | 0.081 |
| Bias (0.06 − 0.18) | 0.040–0.160 | – |
| Scatter (0.15 − 0.7) | 0.036–0.164 | – |
Notes. The last two rows present the CRPSs provided for sources in two cases: with infinitely precise Dirac PDZs but with a bias in range 0.06 − 0.18; and with absolutely accurate (bias = 0) Gaussian PDZs with a scatter in range 0.15 − 0.7.
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.