1. Searching additional two b jets using DNN

First, you need to make pandas arrays to train and evaluate. You can do it using runAna.py in deepAna directory.

In runAna.py, there are several options.

array # If you turn on this option and run, runAna.py make arrays.
array_test # For test. If you turn on this, runAna.py make only test set.
array_train # If you turn on this option with array, runAna.py make training input.
array_syst # If you turn on this option with array, runAna.py make systematic array sets.

Run as follows,

$ python runAna.py

If you make pandas arrays successfully, then you can make model file using model.py

Run as follows,

$ python model.py

Then, you can make histogram using runAna.py with analysis option.

analysis # If you turn on this option, runAna.py make histogram root files.
ana_test # For test.
ana_syst # For systematics

Run as follows,

$ python runAna.py

Then, final root files are made.

2. Differential cross section measurement.

All the codes for measurement are in ttbbDiffXsec directory.

You need to made acceptance distribution first.

$ root -l -b -q makeCriteria.C

Then, stability, purity and acceptance plots are made in output/post directory. And then you can run runUnfold.py. If you run this, you can get unfolding results.

$ python runUnfold.py

you can control unfolding option in ttbbDiffXsec.C

3. Control plots

If you want to make control plots, you can get this by running runAnalysis.py