Feldman-Cousins method to set confidence interval was introduced in Phys. Rev. D 57, 3873 (1998) - Unified approach to the classical statistical analysis of small signals. I want to reproduce figures on the paper for my practice.

In this repository I will follow the notation same as F. and C. In section I they say:

We adopt the following notation: the subscript $t$ on a parameter means the unknown true value; the subscript $0$ means a particlue measured value obtained by an experiment. Thus, for exmample, $\mu$ is a parameter whose true value $\mu_t$ is unknown; $n_0$ is the particluar result of an experiment which measures the number of events, $n$. For most of our discussion, we use for illustration 90% confidence intervals on a single parameter $\mu$. The C.L. is more generally called $\alpha$.

See fig2and3.md

• Original Paper on PDD
• Paper on arXiv
• 三重大奥村さんによる信頼区間の解説