Publication View

A TOOLKIT FOR STATISTICAL COMPARISON OF DATA DISTRIBUTIONS (2008)

Abstract
A typical problem associated to Monte Carlo developments and application consists in the validation of the simulation models and results against experimental data. A novel software toolkit has been developed encompassing an ample variety of statistical algorithms for the comparison of data distributions, such as Monte Carlo simulations and experimental data. The toolkit contains a variety of goodness-of-fit tests, from chi-squared to Kolmogorov-Smirnov, to less known, but generally much more powerful tests such as Anderson-Darling, Cramer-von Mises, Kuiper, Tiku, etc. Thanks to the component-based design and the usage of the standard AIDA interfaces, this tool can be used by other data analysis systems or integrated in experimental software frameworks. We present the architecture of the system, the statistics methods implemented and some results of its applications to the comparison of Geant4 simulations with respect to experimental data.

Publication details
Download http://citeseerx.ist.psu.edu/viewdoc/summary?doi=?doi=10.1.1.133.8879
Source http://www.ge.infn.it/geant4/papers/2005/mc2005/statistics.pdf
Contributors CiteSeerX
Repository CiteSeerX - Scientific Literature Digital Library and Search Engine (United States)
Keywords Key Words, Statistics, Goodness-of-Fit tests, Distributions Comparison, Data Analysis, Software
Type text
Language English