Classification of Higgs boson decays using machine learning. Project for the "Tandem Project" activity at Master degree in Physics.

Table of contents

• Introduction
• Repository diagram structure
• Documentation
• Unofficial paper
• Credits
  • Project leaders
  • Other contributors

Introduction

Study of the Higgs boson Yukawa coupling to tau leptons using the 2012 ATLAS Run-2 dataset. Particular focus is dedicated to the usage of machine learning classification algorithms to classify the Higgs decay channel H to tautau as signal with respect to the other background processes.

For the classification have been considered the cases in which there are 0,1 or 2 jets in the final state.

This classification has been performed on the free dataset from the Higgs Boson Challenge (dataset), that contains data related to the case in which we have in the final state a tau that decays hadronically and the other one that decays leptonically (data for same leptonic or hadronic decays of the tau are omitted).

Analysis scripts are located into the python folder, while Jupyter Notebooks examples are located into the jupyter folder. The purpose of this latter is to show interactively the various analysis passages.

The software is and will stay free, but if you want to support me with a donation it would be really appreciated!

Repository diagram structure

higgs-decay-classification/ ├── doc/ │ ├── PDF_dataset.pdf │ ├── background_explanation.md │ ├── run_the_code.md │ ├── utils.md │ ├── CREDITS.md │ ├── CONTRIBUTING.md ├── img/ ├── scripts/ │ ├── jupyter/ │ │ ├── analysis.ipynb │ │ ├── plots.ipynb │ ├── python/ │ │ ├── analysis.py │ │ ├── plots.py ├── utils/ │ ├── AMS_functions.py │ ├── Make_model.py │ ├── Plot_distributions.py │ ├── Splitting_function.py │ ├── Splitting_jets_function.py │── README.md │── LICENSE │── CITATION.cff │── .gitignore │── .gitattributes

Documentation

List of documentation from the doc folder:

• Background explanation: contains a detailed background explanation of the analysis.
• How to run the code: contains information about how to run the code on your device.
• Utils explanation: contains information about the functions defined for the main program.
• Pdf of the challenge: it is a pdf containing information about the dataset and the challenge.
• Contributing: contains information about how to contribute to this project.

Final results:

Final weighted distribution:

Unofficial paper

An unofficial paper has been produced within this analysis. It has been presented at the 2020 ISHEP school through a small presentation.

This paper can be accessed here.

Credits

Project leaders

Gianluca Bianco

FloMau

Other contributors

Mi. Lia.