Sensitivity optimization of multichannel searches for new signals

TL;DR

This paper extends the frequentist sensitivity definition to multichannel Poisson problems, providing a closed-form formula for complex searches involving multiple signals and backgrounds.

Contribution

It generalizes the sensitivity calculation to multichannel scenarios, offering explicit formulas for Gaussian signals over flat backgrounds.

Findings

Derived a closed-form sensitivity formula for multichannel Poisson problems.

Provided an explicit solution for Gaussian signals on flat backgrounds.

Enhanced the applicability of sensitivity optimization in complex searches.

Abstract

The frequentist definition of sensitivity of a search for new phenomena proposed in arXiv:0308063 has been utilized in a number of published experimental searches. In most cases, the simple approximate formula for the common problem of Poisson counts with background has been deemed adequate for the purpose. There are however many problems nowadays in which more complex analysis is required, involving multiple channels. In this article, the same approach of arXiv:0308063 is applied to a multichannel Poisson problem, and a convenient formula is derived in closed form, generalizing the known result for the simple counting experiment. An explicit solution is also derived for the common case of a search for a Gaussian signal superimposed over a flat background.