DNN-based novel reconstruction and classification approaches in the search for LQs → bτ. Work is submitted as a CMS Experiment Internal Analysis Note-2020 📄.
Data is generated from official CMSSW 10.2.10 Monte Carlo (MC) Simulator with the data sample (2016) of pp collisions ❇️ at a √s = 13 TeV
python 3.6.8
ROOT + PyROOT 6.2
Tensorflow 1.1.4
Keras > 2.0
All other standard libraries like pandas, numpy etc are required. Detailed list and installation in server using anaconda environment is provided here: https://github.com/kschweiger/TF4ttHFH/blob/master/setup_t3PSI.sh
To train the DeepLQ model, choose the type of model from Regression or Classification and execute following. Remenber to change the input data files in each Model_filename.py to actual location of your dataset
python3 Model_filename.py