EMC SPLITOFF STUDIES
Aim is to devise a measure (based on calorimeter cluster quantities)
that will be used to distinguish good photon clusters from
splitoffs. Choosing and using a cut on this will be the
responsibility of whoever is doing the analysis.
Wednesday talks are here ,
and yet again here,
Talk at the July 2004 Collaboration meeting showing the idea and that it works pretty well.
A followup wednesday talk is here.
By now, the discriminant is not looking so much for splitoffs as distinguishing between photons and non-photons. This is the latest talk, given on 3rd November.
Shortly after this, a *small* Beta coding error was noticed which, when fixed, improved the NN performance slightly. Comparaison can be found here.