Including systematic uncertainties in confidence interval construction for Poisson statistics
By J. Conrad, O. Botner, A. Hallgren, C. D. L. Heros. 2002
194 citations recorded by the source corpus.
A search for cosmic point sources of neutrinos has been performed using data taken in 1999 with the AMANDA-B10 neutrino telescope. This work describes methods for signal and background separation and the statistical analysis of the final data sample. In particular, the multivariate method Support Vector Machines has been applied to achieve good background rejection while at the same time retaining high signal efficiency. A grid search covering the complete northern hemisphere revealed no statistical significant excess of events over the expected background from mis-reconstructed cosmic ray induced muons and muons induced by atmospheric neutrinos. Thus, no cosmic point sources of neutrinos have been detected. Upper limits on the neutrino flux for each cell of the grid are presented.Twenty potential sources of neutrinos chosen among three classes of astronomical objects (Blazars, Super Nova Remnants and Microquasars) have been preselected. Upper Limits on the flux of cosmic neutrinos from those are presented.The presence of systematic uncertainties makes the calculation of confidence limits an intricate problem. A method is presented which makes it possible to include these uncertainties into the frequentist construction of confidence intervals. Statistical properties of the presented method have been studied.
Subfields: Nuclear and High Energy Physics
View Paper: https://www.semanticscholar.org/paper/29d32fb98962a767afbd9f2188da0151e296c658
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
Citations in the corpus
- Limits and confidence intervals in the presence of nuisance parameters (2004, background): “over the nuisance parameters, see for example Cousins and Highland [12] and Conrad et. al. [ 13 ].”
- Dark matter direct-detection experiments (2015, unknown): “worth mentioning that other methods exist which include systematic uncertainties in the confidence interval construction [210][211].”