Multivariate analysis techniques
The end products of experimental data analysis are designed to be simple and easy to understand: hypothesis tests and measurements of parameters. But, the experimental data themselves are voluminous and complex. Furthermore, in modern collider experiments, many petabytes of data must be processed in search of rare new processes which occur together with much more copious background processes that are of less interest to the task at hand. The systematic uncertainties on the background may be larger than the expected signal in many cases. The statistical power of an analysis and its sensitivity to systematic uncertainty can therefore usually both be improved by separating signal events from background events with higher efficiency and purity.
- Research Organization:
- Fermi National Accelerator Lab. (FNAL), Batavia, IL (United States)
- Sponsoring Organization:
- USDOE Office of Science (SC), High Energy Physics (HEP)
- DOE Contract Number:
- AC02-07CH11359
- OSTI ID:
- 1408510
- Report Number(s):
- FERMILAB-PUB-16-733-E; 1497147
- Resource Relation:
- Related Information: Appendix C in the book, "Discovery of the Higgs Boson", A. Nisait and V. Sharma, eds. World Scientific, 2016; ISBN 9814425877
- Country of Publication:
- United States
- Language:
- English
Similar Records
Foundations of Molecular 'Isotomics'
Theoretical Studies in Heterogenous Catalysis: Towards a Rational Design of Novel Catalysts for Hydrodesulfurization and Hydrogen Production