← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6578.d · 生成模型:GPT-6

结构线性与回归线性有别

理解等级:高级

观测条件均值E[Y|X]的线性是统计约束;结构方程Y:=aX+U的线性则限制干预与跨世界机制。两者都不完整显示在DAG上,但不能因此把所有图外参数假设都归为反事实假设。

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

      to think about this is that any causal assumption you cannot represent in the structure of the DAG is, by process of elimination, a level 3 assumption. For example, suppose your DAG has the edge X →Y. Further, you believe the causal relationship between X and Y is naturally linear. You can’t “see” linearity on the DAG structure, so linearity is a level 3 assumption.

    参考来源

    - 位置:《Causal AI》第10章《识别与因果层级》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-10.xhtml) - 结论依据:摘录中关于结构线性与回归线性有别的定义、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:to think about this is that any causal assumption you cannot represent in the structure of the DAG is, by process of elimination, a level 3 assumption. For example, suppose your DAG has the edge X →Y. Further, you believe the causal relationship between X and Y is naturally linear. You can’t “see” linearity on the DAG structure, so linearity is a level 3 assumption.