Code to train various classifiers, aiming to optimise categorisation for the Higgs to diphoton analysis at Stage 1 of the Simplified Template Cross-section (STXS) framework.

This is all desiged to run on the IC LX machines, in a CMSSW 10 environment. Below are some instructions which should get you up and running. Starting from a clean area:

cmsrel CMSSW_10_2_0
cd CMSSW_10_2_0/src/
cmsenv
git cms-init
git clone https://github.com/edjtscott/Stage1categorisation.git

If you don't have a github account yet, it's probably worth creating one and forking this repository so you can play with it and update as needed. Once you do, then you'd add something like

git remote add yourusername https://github.com/yourusername/Stage1categorisation.git
git checkout -b yourbranchname

to create a new branch. As you can see, there is currently just one directory, TwoStep. It is so-called because the code there does the Stage 1 classification in two steps: assigning the signal bin first, then doing background discrimination. Since doing both simultaneously in one classifier would also be possible, a OneStep directory might be handy at some point. See the readme in TwoStep for explanations of what each script does.