[论文解读] Negatively Biased Relevant Subsets Induced by the Most-Powerful One-Sided Upper Confidence Limits for a Bounded Physical Parameter

R. Cousins|arXiv (Cornell University)|Sep 9, 2011

Probabilistic and Robust Engineering Design参考文献 13被引用 6

一句话总结

本文批判了高能物理中用于计算单侧上限的“原始对角线法”，指出该方法会导致负向偏倚的相关子集，从而破坏事后覆盖概率。文章认为，即使经过修改的版本（如带功率约束或截断的版本）也无法解决核心问题，主张采用Feldman-Cousins双侧似然比排序方法作为更优替代方案，因其具备更佳的条件覆盖特性。

ABSTRACT

Suppose an observable x is the measured value (negative or non-negative) of a true mean mu (physically non-negative) in an experiment with a Gaussian resolution function with known fixed rms deviation s. The most powerful one-sided upper confidence limit at 95% C.L. is UL = x+1.64s, which I refer to as the "original diagonal line". Perceived problems in HEP with small or non-physical upper limits for x<0 historically led, for example, to substitution of max(0,x) for x, and eventually to abandonment in the Particle Data Group's Review of Particle Physics of this diagonal line relationship between UL and x. Recently Cowan, Cranmer, Gross, and Vitells (CCGV) have advocated a concept of "power constraint" that when applied to this problem yields variants of diagonal line, including UL = max(-1,x)+1.64s. Thus it is timely to consider again what is problematic about the original diagonal line, and whether or not modifications cure these defects. In a 2002 Comment, statistician Leon Jay Gleser pointed to the literature on recognizable and relevant subsets. For upper limits given by the original diagonal line, the sample space for x has recognizable relevant subsets in which the quoted 95% C.L. is known to be negatively biased (anti-conservative) by a finite amount for all values of mu. This issue is at the heart of a dispute between Jerzy Neyman and Sir Ronald Fisher over fifty years ago, the crux of which is the relevance of pre-data coverage probabilities when making post-data inferences. The literature describes illuminating connections to Bayesian statistics as well. Methods such as that advocated by CCGV have 100% unconditional coverage for certain values of mu and hence formally evade the traditional criteria for negatively biased relevant subsets; I argue that concerns remain. Comparison with frequentist intervals advocated by Feldman and Cousins also sheds light on the issues.

研究动机与目标

• 分析高能物理中广泛使用的'原始对角线法'在计算单侧上限时存在的统计缺陷。
• 研究为何诸如在零处截断或施加功率约束的修改版本，仍无法解决相关子集中负向偏倚的问题。
• 评估频率学派置信区间与事前及事后覆盖评估的一致性，尤其在Neyman-Fisher论争的背景下。
• 主张采用Feldman-Cousins双侧似然比排序方法作为更可靠的替代方案，以避免相关子集中严重偏倚。
• 通过相关子集与覆盖偏倚的视角，阐明频率学派置信区间与贝叶斯方法之间的联系。

提出的方法

• 分析原始对角线法的置信带构造，其中 μ_UL = x + 1.64σ，表明该方法会诱导出可识别的相关子集，且对所有 μ 值均存在已知的负向偏倚。
• 应用统计文献中关于相关子集的概念，特别是Gleser的工作，证明对于所有 μ，95%置信水平在某些数据区域中呈现反保守性。
• 将原始方法与修改版本进行比较：在零处截断（μ_UL = max(0, x+1.64σ)）和功率约束极限（例如 μ_UL = max(−1, x) + 1.64σ），表明它们仍存在覆盖问题。
• 评估Feldman-Cousins方法，该方法采用双侧似然比排序，表明其提供了更好的条件覆盖，并避免了最坏情况下的偏倚。
• 使用Neyman-Pearson框架对比事前无条件覆盖与事后条件覆盖，凸显对置信水平解释中的张力。
• 回顾与贝叶斯统计的联系，特别是先验如何产生具有类似频率学派区间覆盖特性的可信区间，暗示两种范式之间存在桥梁。

实验结果

研究问题

• RQ1 为何原始对角线法在上限计算中会导致事后推断中相关子集的负向偏倚？
• RQ2 对角线法的功率约束或截断修改版本是否能消除相关子集中覆盖偏倚的问题？
• RQ3 Feldman-Cousins方法在条件覆盖与无条件覆盖方面的覆盖特性，与对角线法及其变体相比如何？
• RQ4 贝叶斯方法在多大程度上能够启发或解决频率学派置信区间中的覆盖偏倚问题？
• RQ5 能否构建一种通用方法，确保在所有相关子集中实现良好覆盖，同时保持高统计功效与可解释性？

主要发现

• 原始对角线法（μ_UL = x + 1.64σ）会诱导出可识别的相关子集，其中对所有 μ 值，95%置信水平均系统性地呈现反保守性。
• 即使经过修改，如 μ_UL = max(0, x+1.64σ) 或 μ_UL = max(−1, x) + 1.64σ，仍无法完全消除相关子集中负向偏倚的问题。
• Feldman-Cousins双侧似然比排序方法提供了更好的条件覆盖，并避免了最坏情况下的偏倚，因此是更可靠的替代方案。
• 问题根源在于事前无条件覆盖与事后条件覆盖之间的根本性冲突，而Neyman-Pearson框架无法解决此问题。
• 关于相关子集的文献揭示了与贝叶斯统计的深层联系，表明能产生良好频率学派覆盖效果的先验，或可作为务实的前进路径。
• 目前尚无通用方法可同时确保所有相关子集及超集中的良好覆盖，凸显了现有频率学派方法的局限性。