
The Library
A new maximum-likelihood method for template fits
Tools
Dembinski, Hans and Abdelmotteleb, A. S. W. (2022) A new maximum-likelihood method for template fits. The European Physical Journal C, 82 (11). 1043. doi:10.1140/epjc/s10052-022-11019-z ISSN 1434-6052.
|
PDF
WRAP-new-maximum-likelihood-method-template-fits-22.pdf - Published Version - Requires a PDF viewer. Available under License Creative Commons Attribution 4.0. Download (957Kb) | Preview |
Official URL: https://doi.org/10.1140/epjc/s10052-022-11019-z
Abstract
A common statistical problem in particle physics is to extract the number of samples which originate from a statistical process in an ensemble containing a mix of several contributing processes. The probability density function of each process is usually not exactly known. Barlow and Beeston found an exact likelihood for the problem of fitting binned templates obtained from Monte-Carlo simulation to binned data, which propagates the uncertainty of the templates into the result. Solving the exact likelihood is technically challenging, however. The original paper also did not provide a way to use weighted simulation samples with varying weights. Other papers have introduced alternative likelihoods to address these points. In this paper, a new approximate likelihood is derived from the exact Barlow–Beeston one. The new likelihood is generalized to fits of weighted templates to weighted data. The performance of the new likelihood is evaluated based on toy examples. The performance is excellent – point estimates have small bias and confidence intervals have good coverage – and is comparable to the exact Barlow–Beeston likelihood when the templates are not weighted. The new likelihood evaluates faster than the Barlow–Beeston one when the number of bins is large.
Item Type: | Journal Article | ||||||
---|---|---|---|---|---|---|---|
Divisions: | Faculty of Science, Engineering and Medicine > Science > Physics | ||||||
SWORD Depositor: | Library Publications Router | ||||||
Journal or Publication Title: | The European Physical Journal C | ||||||
Publisher: | Springer Berlin Heidelberg | ||||||
ISSN: | 1434-6052 | ||||||
Official Date: | 17 November 2022 | ||||||
Dates: |
|
||||||
Volume: | 82 | ||||||
Number: | 11 | ||||||
Article Number: | 1043 | ||||||
DOI: | 10.1140/epjc/s10052-022-11019-z | ||||||
Status: | Peer Reviewed | ||||||
Publication Status: | Published | ||||||
Reuse Statement (publisher, data, author rights): | ** From Springer Nature via Jisc Publications Router ** History: received 27-07-2022; collection 11-2022; accepted 07-11-2022; registration 08-11-2022; pub-electronic 17-11-2022; online 17-11-2022. ** Licence for this article: http://creativecommons.org/licenses/by/4.0/ ** Acknowledgements: Acknowledgements: We thank Roger Barlow, Carlos Argüelles, Austin Schneider, and Tianlu Yuan for valuable discussions in preparation of this work. | ||||||
Access rights to Published version: | Open Access (Creative Commons) | ||||||
Date of first compliant deposit: | 9 December 2022 | ||||||
Date of first compliant Open Access: | 9 December 2022 | ||||||
Related URLs: |
Request changes or add full text files to a record
Repository staff actions (login required)
![]() |
View Item |
Downloads
Downloads per month over past year