Table C.2
Performance metrics as percentages (precision and recall for the pre-mergers and post-mergers) of the different methods on the TNG-test set per redshift bin, for the multi-class classification task. The post-merger class includes the ongoing-mergers. The merger classes or stages are determined by the merger trees from the TNG simulation. The best performance in each metric is highlighted in bold.
0.1 < ɀ < 0.31 | 0.31 < ɀ < 0.52 | 0.52 < ɀ < 0.76 | 0.76 < ɀ < 1.0 | |||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Pre-merger | Post-merger | Pre-merger | Post-merger | Pre-merger | Post-merger | Pre-merger | Post-merger | |||||||||
Methods | P | R | P | R | P | R | P | R | P | R | P | R | P | R | P | R |
M-1 (RF) | 62.9 | 48.5 | 56.2 | 68.0 | 63.2 | 45.9 | 57.1 | 65.9 | 58.7 | 40.1 | 53.7 | 65.0 | 53.8 | 36.8 | 51.3 | 59.5 |
M-2 (Swin) | 76.6 | 66.3 | 61.0 | 73.3 | 75.8 | 69.4 | 59.4 | 70.9 | 70.5 | 66.5 | 59.4 | 64.0 | 65.6 | 60.8 | 61.1 | 52.8 |
M-3 (Zoobot) | 69.5 | 76.3 | 85.6 | 42.1 | 68.6 | 76.8 | 83.0 | 43.8 | 65.0 | 75.3 | 77.5 | 35.8 | 59.0 | 68.9 | 77.6 | 30.7 |
M-4 (CNN1) | 68.8 | 29.8 | 41.1 | 85.2 | 67.6 | 23.3 | 41.3 | 83.4 | 59.1 | 8.9 | 41.5 | 73.5 | 60.5 | 12.4 | 43.6 | 66.5 |
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.