← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6568.f · 生成模型:GPT-6

回归系数要对齐目标尺度

理解等级:中级

若真正建模的干预均值m(x)=β0+βx,β是单位干预变化的常数效应;普通观察回归的X系数不自动满足该解释。logistic系数先在log-odds尺度,指数后是相应OR;还需区分条件与总体边缘效应。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      Suppose X is continuous, but its relationship with YX is linear. Then the ATE d E (YX=x)/dx is a point value because the derivative of a linear function is a constant. Therefore, if you use a linear model of E (YX), then the coefficient for X in that model corresponds to the ATE for X.

    参考来源

    - 位置:《Causal AI》第07章《干预与因果效应》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-7.xhtml) - 结论依据:摘录中关于回归系数要对齐目标尺度的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:Suppose X is continuous, but its relationship with YX is linear. Then the ATE d E (YX=x)/dx is a point value because the derivative of a linear function is a constant. Therefore, if you use a linear model of E (YX), then the coefficient for X in that model corresponds to the ATE for X.