BEXCO - Room F(201/202/203/204)
[GA010] Research on statistical methods for signal significance estimation in ground-based gamma-ray detectors
Searching for new astrophysical sources is an important physical goal of current and nearly further ground-based gamma-ray detectors, such as HAWC, LHAASO and CTA. The gamma ray signal appears as an excess number of events. The significance of the excess is estimated using statistical method and is used to determine whether to accept the signal or not. Up to now, four main statistical methods, including the classical Li-Ma method, weighted method, likelihood method and chi-squared method, have been used in the majority of the EAS experiments, such as Tibet ASγ, MILAGRO, ARGO-YBJ and HAWC. While most these methods are not well tested for the accuracy in previous works. In this work, we develop a standard to test the accuracy of different methods basing on different experimental conditions. We find some problems for some methods, which have been already used in experiment data analysis. Finally, we obtain the limitation of different methods and provide application conditions for different methods, respectively. Additionally, we also test the superiority of different statistical methods and provide our best recommendations for different physical analysis.