arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2606.23116stat.ME
收藏

广义线性模型中的直接与间接歧视

Direct and Indirect Discrimination in Generalized Linear Models

Milliman R&D(密理曼研发部);Université du Québec à Montréal(魁北克大学蒙特利尔分校);Kyoto University(京都大学);LSAF, Université Claude Bernard Lyon 1(里昂第一大学 LSAF)

Bertille Tierny, Arthur Charpentier, François Hu

AI总结:

针对广义线性模型预测中的群体差异,提出基于矩的分解框架,区分直接均值、间接均值、交互、结构及曲率耦合/放大成分,并应用于医疗支出数据。

摘要:

广义线性模型是二元风险、索赔频率、利用率和成本相关结果的精算建模核心。然而,公平性诊断通常依赖于线性模型直觉,尽管GLM预测是通过将潜在得分通过非线性逆链接函数转换得到的。我们开发了一个基于矩的分解框架,用于诊断拟合GLM预测中的群体差异。在精确的线性-高斯基准中,分布性人口统计平等的Wasserstein重心准则简化为两矩准则,并分解为直接均值、间接均值、交互和结构成分。对于GLM,我们区分经验输出尺度准则$U_2(f)$、组内代理$\widetilde U_2(f)$和主导分解$D_1(f)$。这一主导项保留了四个线性通道,并增加了由逆链接函数引起的两个曲率成分:曲率耦合和曲率放大。我们推导了逻辑、泊松和Tweedie规格的显式公式,并在医疗支出调查数据上说明了该诊断。该框架不是歧视的法律测试,也不是线性-高斯情况外分布性平等的完整刻画。它是一个易于处理的精算诊断工具,用于识别拟合预测差异是否源于显式敏感效应、代理介导的协变量轮廓、协方差结构差异或非线性链接效应。

英文摘要:

Generalized linear models are central to actuarial modelling of binary risk, claim frequency, utilization, and cost-related outcomes. Yet fairness diagnostics often rely on linear-model intuitions, although GLM predictions are obtained by transporting a latent score through a nonlinear inverse link. We develop a moment-based decomposition framework for diagnosing group disparities in fitted GLM predictions. In an exact linear-Gaussian benchmark, the Wasserstein barycentric criterion for distributional demographic-parity violation reduces to a two-moment criterion and decomposes into direct mean, indirect mean, interaction, and structural components. For GLMs, we distinguish the empirical output-scale criterion $U_2(f)$, a within-group proxy $\widetilde U_2(f)$, and a leading decomposition $D_1(f)$. This leading term preserves the four linear channels and adds two curvature components induced by the inverse link: curvature coupling and curvature amplification. We derive explicit formulas for logistic, Poisson, and Tweedie specifications and illustrate the diagnostic on medical-expenditure survey data. The framework is not a legal test of discrimination, nor a full characterization of distributional parity outside the linear-Gaussian case. It is a tractable actuarial diagnostic for identifying whether fitted prediction disparities arise from explicit sensitive effects, proxy-mediated covariate profiles, covariance-structure differences, or nonlinear link effects.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑