Reducing bias in the analysis of counting statistics data

被引：3

作者：

Hammersley, AP ^{[1
]}

Antoniadis, A ^{[1
]}

机构：

[1] UNIV GRENOBLE,IMAG,GRENOBLE,FRANCE

来源：

NUCLEAR INSTRUMENTS & METHODS IN PHYSICS RESEARCH SECTION A-ACCELERATORS SPECTROMETERS DETECTORS AND ASSOCIATED EQUIPMENT | 1997年 / 394卷 / 1-2期

关键词：

data analysis; counting statistics; Poisson statistics; least squares fitting; maximum likelihood methods; model fitting;

D O I：

10.1016/S0168-9002(97)00668-2

中图分类号：

TH7 [仪器、仪表];

学科分类号：

0804 ; 080401 ; 081102 ;

摘要：

In the analysis of counting statistics data it is common practice to estimate the variance of the measured data points as the data points themselves. This practice introduces a bias into the results of further analysis which may be significant, and under certain circumstances lead to false conclusions. In the case of normal weighted least squares fitting this bias is quantified and methods to avoid it are proposed.

引用

页码：219 / 224

页数：6