Table 3.
Best performing models for the AGN-galaxy classification.
| Model | Fβ | MCC | Precision | Recall | Rank |
|---|---|---|---|---|---|
| (× 100) | (× 100) | (× 100) | (× 100) | ||
| CatBoost | 95.70 ± 0.28 | 92.46 ± 0.48 | 95.45 ± 0.32 | 95.91 ± 0.37 | 1.00 |
| XGBoost | 95.67 ± 0.27 | 92.40 ± 0.48 | 95.41 ± 0.39 | 95.88 ± 0.34 | 2.00 |
| RF | 95.52 ± 0.36 | 92.14 ± 0.63 | 95.28 ± 0.46 | 95.71 ± 0.40 | 3.00 |
| ET | 95.40 ± 0.40 | 91.94 ± 0.69 | 95.13 ± 0.43 | 95.63 ± 0.47 | 4.00 |
| GBC | 95.26 ± 0.31 | 91.66 ± 0.54 | 94.82 ± 0.41 | 95.63 ± 0.35 | 5.00 |
Notes. Metrics obtained using the default probability threshold of 0.5. Algorithms are sorted by decreasing recall values. For display purposes, all metrics have been multiplied by 100. Uncertainties show standard deviation of metrics obtained across all 10 training folds (cf. Sect. 3.3).
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.