Fast Conservative Monte Carlo Confidence Sets
提出一种通用方法，通过反转蒙特卡洛检验来构造实值和多维参数的保守置信集，适用于多种随机化方案，并提供了开源实现。
We present a general approach to construct confidence sets for real-valued and multidimensional parameters for a broad range of randomization schemes by inverting Monte Carlo tests. The approach exploits two facts: (i) there are Monte Carlo tests that are conservative despite relying on simulation, and (ii) since the coverage probability of confidence sets depends only on the significance level of the test of the true null, every null can be tested using the same Monte Carlo sample. When the parameter is real-valued and the p-value is quasiconcave in the parameter with the data and Monte Carlo sample held fixed, conservative confidence sets can be constructed in O(n) time, where n is the number of data. The values of some test statistics for different Monte Carlo samples and parameter values have a simple relationship that makes more savings possible. We provide open-source Python and R implementations, available online.