Improved probability method for estimating signal in the presence of background
Improved probability method for estimating signal in the presence of background
复制标题
存在背景时估计信号的改进概率方法
DOI：
10.1103/physrevd.60.053009
复制
发表时间：
1998
影响因子：
5
通讯作者：
M. Woodroofe
中科院分区：
文献类型：
--
作者：
B. Roe;M. Woodroofe
本文对Feldman-Schumins 1估计有背景时信号计数的方法提出了改进意见。该方法集中于寻找信号的基本信息，而忽略背景的无关信息。找到了一个合适的方法，它使用的条件是，所获得的回退事件的数量不超过所获得的事件的总数。探讨了几种替代方法。1. Feldman和Feldins，1在最近的一篇文章中，在解决两个长期存在的问题方面取得了重大进展，这两个问题涉及使用置信水平来估计来自数据的概率。其中第一个是消除在检查数据后，在使用置信区间或置信限之间做出决定时出现的偏差。第二个是当实验结果产生的估计接近或超过感兴趣参数的已知界限时，找到一个置信区间。Feld-man和Eldins的方法在下文中称为统一艾德方法，并在第2节中进行了描述。在本文中，我们认为，统一的艾德的方法并不十分足够的津贴为已知的界限，并建议修改。该修改是根据卡门2号数据进行的
ABSTRACT A suggestion is made for improving the Feldman-Cousins 1 method of estimating sig-nal counts in the presence of background. The method concentrates on finding essentialinformation about the signal and ignoring extraneous information about background. Anappropriate method is found which uses the condition that that the number of backgroundevents obtained does not exceed the total number of events obtained. Several alternativeapproaches are explored. 1. Introduction Feldman and Cousins, 1 in a recent article, have made major advances towards solv-ing two long-standing problems concerning the use of confidence levels for estimating aparameter from data. The first of these is eliminating the bias that occurs when onedecides between using a confidence interval or a confidence bound, after examining thedata. The second is finding a confidence interval when the experimental result producesestimators that are close to or past known bounds for the parameters of interest. Feld-man and Cousins’ method is called the unified approach below and is described in Section2. In the present paper we argue that the unified approach does not make quite enoughof an allowance for the known bounds and suggest a modification. The modification isillustrated with the KARMEN 2 Data