Open Access

Fig. 5

image

Download original image

Relative change of the signal efficiency ϵsig during the iterative removal of the least important variable (from bottom to top). The signal efficiency for each training is compared to the signal efficiency of the reference training ϵsigHillasOnly $_{{\rm{sig}}}^{{\rm{HillasOnly }}}$ (the current training with the 6 Hillas variables) for a background rejection of 95%. The black error bars show the difference in the signal efficiency for simulations with nominal and high NSB values, which exceeds the difference present in the reference training. No error bar corresponds to a difference which is smaller than in the reference training. Variables below the dashed red line are not selected for the final training.

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.