Here we have three notebooks that are independent of each other, which will teach you different things about statistical inference with pyhf and convenience packages such as cabinetry.
The different notebooks include the following:
- 01-histogram-fits: This introduces the basics of pyhf, how the model building works and how to include uncertainties. We encourage you to start with this notebook.
-
02-B2Kpi: This notebook is a realistic example of how to build a statistical model for the
$B^+ \to K^+ \pi^0$ decay. We use reconstructed MC in combination with cabinetry to build our pyhf model. Additionally, tracking efficiency and PID systematics are included in the model. The 02-pid-weights notebook can be studied as an extension, but is not required (access to KEKCC is necessary to access the ntuples to run this notebook). - 03-hypothesis-testing: This notebook goes beyond the basics in pyhf, and introduces advanced methods of statistical inference, such as hypothesis testing on a very simple model.