A repository combining the full workflow of PAIReD tagger development.

PAIReD jets are new types of jets for improved reconstruction of hadronic decays of heavy objects such as the Higgs boson. This repository deals with the creation of a PAIReD tagger to classify these newly defined jets. From the production of PAIReD data sets to the training of a neural network and its testing, everything will be covered in this repository.

The individual steps of the workflow required to obtain a PAIReD tagger are described here in several chapters:

1. Production of data sets
2. Training a PAIReD tagger
3. Evaluation of tagger performance

For installing this framework, go to the directory on your machine where you want to install it. Make sure that it is a reasonable place. Then, clone this repository to your current directory by running:

git clone https://github.com/JanGerritSchulz/PAIReD_jet_tagging.git

This framework enables the production of PAIReD jet files. Those are ROOT files containing a list of PAIReD jets, meaning each entry is one PAIReD jet. A summary on the stored variables and MC information can be found in the PAIReD file content overview. This includes the definition of the labels and some selections on the seed AK4 jets used for building the PAIReD jets.

You can find a number of pre-trained PAIReD taggers in PAIReD_Tagger_Training/trained-models.