Free Access

Table 5.

Area under the ROC curve (AUC) for networks trained on different planed datasets of the m12i LSR0 catalog, and using only 5D kinematics.

Data AUC
Using full set 0.96
Removed ϕ 0.97
Removed R 0.94
Removed z 0.90
Removed vz 0.84
Removed vR 0.83
Removed vϕ 0.75

Notes. The extent to which the AUC is lowered when removing information represents how important that variable is in distinguishing accreted stars. For comparison, we also provide the AUC for the original (un-planed) data. Of the simple kinematic variables considered, the disk rotation is the most crucial, lowering the AUC from 0.96 to 0.75. This suggests that, even though the machine is not given enough information to fully reconstruct vϕ, it is still inferring a set of correlations of the input variables that correspond to this quantity.

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.