Speakers
Description
Summary
HistFitter is a software framework for statistical data analysis used widely throughout the ATLAS Collaboration at the LHC. HistFitter interfaces with the statistical tools HistFactory and RooStats to construct parametric models and to perform statistical tests of the data, and extends these tools in four key areas:
-
Programmable framework: HistFitter puts tools from several sources together in a coherent and programmable framework, capable of performing a complete statistical analysis of pre-formatted input data samples.
-
Bookkeeping: HistFitter can perform statistical tests and scan over parameter values of hundreds of signal hypotheses in an organized way from a single user-defined configuration file.
-
Analysis strategy: HistFitter has built-in the concepts of control, signal and validation regions, which are used to constrain, extrapolate and validate data-model predictions across analysis regions. HistFitter also introduces a rigorous treatment of validation regions that is new in high-energy physics.
-
Presentation and interpretation: the HistFitter framework keeps track of data models before and after fits to the measured data, and includes a collection of methods to determine the statistical significance using various statistical tests and to produce tables and plots expressing these results with publication-quality style.
