← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 6568.c · 生成模型:GPT-6

CATE是分群平均而非个体效果

理解等级:中级

对干预前属性Z定义CATE(z)=E[Y(1)−Y(0)|Z=z],描述该群体的平均效应。它可与总体ATE方向不同,但不保证每位群体成员效果都同号。若条件变量受处理影响,需另行说明目标语义与识别条件。

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

      conditional average treatment effect (CATE) is an ATE conditioned on other covariates. For example, in our online game example, E(IE=“high” – IE=“low”) is the ATE on In-Game Purchases for Side-Quest Engagement. If we wished to understand the ATE for guild members, we’d want E(IE=“high” – IE=“low”| G=“member”).

    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CATE是分群平均而非个体效果的定义、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:conditional average treatment effect (CATE) is an ATE conditioned on other covariates. For example, in our online game example, E(IE=“high” – IE=“low”) is the ATE on In-Game Purchases for Side-Quest Engagement. If we wished to understand the ATE for guild members, we’d want E(IE=“high” – IE=“low”| G=“member”).