HistFitter: a flexible framework for statistical data analysis

21st International Conference on Computing in High Energy and Nuclear Physics

物理学;计算机科学

Besjes, G.J.^1 ; Baak, M.^2 ; Côté, D.^3 ; Koutsman, A.^4 ; Lorenz, J.M.^5,6 ; Short, D.^7

Niels Bohr Institute, University of Copenhagen, Blegdamsvej 17, Copenhagen

2100, Denmark^1

CERN, Route de Meyrin 385, Meyrin

1217, Switzerland^2

University of Texas at Arlington, 701 South Nedderman Drive, Arlington

TX

76019, United States^3

TRIUMF, 4004 Wesbrook Mall, Vancouver

BC

2A3, Canada^4

Fakultat für Physik, LMU München, Am Coulombwall 1, Garching

85748, Germany^5

Excellence Cluster Universe, Boltzmannstr. 2, Garching

85748, Germany^6

Department of Physics, Denys Wilkinson Building, University of Oxford, Keble Road, Oxford

OX1 3RH, United Kingdom^7

关键词: De facto standard;    Flexible framework;    Large Hadron Collider;    Multiple tools;    Parametric models;    Proton proton collisions;    Software frameworks;    Statistical data analysis;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/664/7/072004/pdf
DOI  :  10.1088/1742-6596/664/7/072004

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

HistFitter is a software framework for statistical data analysis that has been used extensively in the ATLAS Collaboration to analyze data of proton-proton collisions produced by the Large Hadron Collider at CERN. Most notably, HistFitter has become a de-facto standard in searches for supersymmetric particles since 2012, with some usage for Exotic and Higgs boson physics. HistFitter coherently combines several statistics tools in a programmable and flexible framework that is capable of bookkeeping hundreds of data models under study using thousands of generated input histograms. HistFitter interfaces with the statistics tools HistFactory and RooStats to construct parametric models and to perform statistical tests of the data, and extends these tools in four key areas. The key innovations are to weave the concepts of control, validation and signal regions into the very fabric of HistFitter, and to treat these with rigorous methods. Multiple tools to visualize and interpret the results through a simple configuration interface are also provided.