Abstract
We propose modified frequentist definition for the determination of confidence intervals for the case of Poisson statistics. Namely, we require that \(1 - \beta ' \geqslant \sum\nolimits_{n = 0}^{n = n_{obs} + k} {P\left( {\left. n \right|\lambda } \right)} \geqslant \alpha '\). We show that this definition is equivalent to the Bayesian method with prior π(λ) ∼ λk. We also propose modified frequentist definition for the case of nonzero background.
Similar content being viewed by others
References
As a review, see for example: F. James, Statistical Methods in Experimental Physics, 2nd ed. (World Scientific, 2006).
See, for example: G. D. Agostini, Bayesian Reasoning in Data Analysis, a Critical Introduction. Hackensack (World Scientific, NJ, 2003).
J. Neyman, Philos. Trans. R. Soc. Sect. A (London), 236, 333–380 (1937).
F. Garwood, Biometrica 28, 437 (1936).
W. E. Ricker, J. Amer. Stat. Assoc. 32, 349 (1937).
R. D. Cousins, Am. J. Phys 63, 398 (1995).
T. Junk, Nucl. Instrum. Meth. A 434, 435 (1999).
A. L. Read, Modified Frequentist Analysis of Search Results (the CL s Method): CERN Yellow Report CERN-2000-005, 2000, p. 81.
Author information
Authors and Affiliations
Additional information
The article is published in the original.
Rights and permissions
About this article
Cite this article
Bitioukov, S.I., Krasnikov, N.V. Modified frequentist determination of confidence intervals for poisson distribution. Phys. Part. Nuclei 44, 229–233 (2013). https://doi.org/10.1134/S1063779613020081
Published:
Issue Date:
DOI: https://doi.org/10.1134/S1063779613020081