Machine Learning EPD centrality

Attached to this blog are updates from George Halal.

The EPD presentation can be found at: drupal.star.bnl.gov/STAR/system/files/EPDMeeting_030518_GeorgeHalal.pdf

Looking at the derivative of the ANN tells us how sensitive a given row is to the impact parameter.  A value of 0 means not sensitive, higher numbers indicate more sensitivity. 

sqrt(S_NN) = 54 GeV:
Quarter wheel: drupal.star.bnl.gov/STAR/system/files/MLPDIn_QW54.0.pdf
Full EPD: drupal.star.bnl.gov/STAR/system/files/MLPDIn_54.0.pdf

sqrt(S_NN) = 19.6 GeV: drupal.star.bnl.gov/STAR/system/files/MLPDIn_19.6.pdf

sqrt(S_NN) = 14.5 GeV: drupal.star.bnl.gov/STAR/system/files/MLPDIn_14.5.pdf

sqrt(S_NN) = 11.5 GeV: drupal.star.bnl.gov/STAR/system/files/MLPDIn_11.5.pdf

sqrt(S_NN) = 7.7 GeV: drupal.star.bnl.gov/STAR/system/files/MLPDIn_7.7.pdf