Lars-Erik De Geer1. 1. Preparatory Commission for the Comprehensive, Nuclear-Test-Ban Treaty Organization, Provisional Technical Secretariat, Vienna International Centre, P.O. Box 1200, 1400, Austria. ledg@ctbto.org
Abstract
Currie Hypothesis testing is applied to gamma-ray spectral data, where an optimum part of the peak is used and the background is considered well known from nearby channels. With this, the risk of making Type I errors is about 100 times lower than commonly assumed. A programme, PeakMaker, produces random peaks with given characteristics on the screen and calculations are done to facilitate a full use of Poisson statistics in spectrum analyses. SHORT TECHNICAL NOTE SUMMARY: The Currie decision limit concept applied to spectral data is reinterpreted, which gives better consistency between the selected error risk and the observed error rates. A PeakMaker program is described and the few count problem is analyzed.
Currie Hypothesis testing is applied to gamma-ray spectral data, where an optimum part of the peak is used and the background is considered well known from nearby channels. With this, the risk of making Type I errors is about 100 times lower than commonly assumed. A programme, PeakMaker, produces random peaks with given characteristics on the screen and calculations are done to facilitate a full use of Poisson statistics in spectrum analyses. SHORT TECHNICAL NOTE SUMMARY: The Currie decision limit concept applied to spectral data is reinterpreted, which gives better consistency between the selected error risk and the observed error rates. A PeakMaker program is described and the few count problem is analyzed.
Authors: Maite Jauregui-Osoro; Simona De Robertis; Philip Halsted; Sarah-May Gould; Zilin Yu; Rowena L Paul; Paul K Marsden; Antony D Gee; Andrew Fenwick; Philip J Blower Journal: Nucl Med Commun Date: 2021-09-01 Impact factor: 1.698