Table 1.
Performance of the classification methods in terms of precision, recall, and F1 score.
| Method | Precision | Recall | F1 score | |
|---|---|---|---|---|
| (red/blue) | (red/blue) | (red/blue) | ||
| NUV, u, r, i, z, Y, J, H, K | ||||
| LDA | 0.94/0.99 | 0.84/1.00 | 0.89/0.99 | |
| LDA + SMOTE | 0.69/1.00 | 1.00/0.97 | 0.81/0.98 | |
| XGB | 0.96/1.00 | 0.95/1.00 | 0.95/1.00 | |
| XGB with −99.9 | 0.96/1.00 | 0.93/1.00 | 0.94/1.00 | |
| NUV, r, J, K | ||||
| LDA | 0.90/0.99 | 0.90/0.99 | 0.90/0.99 | |
| LDA + SMOTE | 0.66/1.00 | 1.00/0.96 | 0.80/0.98 | |
| XGB | 0.96/1.00 | 0.95/1.00 | 0.95/1.00 | |
| XGB with −99.9 | 0.95/0.99 | 0.91/1.00 | 0.93/1.00 | |
Notes. The upper block refers to the training with all available rest-frame magnitude bands, the bottom block to the one using only the four with the highest decision impact. Testing set sample size and red/blue ratio: 94236 galaxies with 7.1% red/blue ratio; 98659 with 7.2% red/blue ratio, when missing values are imputed.
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.