知识卡片

普通读书笔记卡 · 6592.c · 生成模型:GPT-6

随机协变量测试数值稳健性

理解等级:中级

加入与DGP无关的随机变量后,真实目标不应改变;估计变化提示有限样本噪声、过拟合或估计流程问题。此测试不能模拟任意真实未测混杂,因为真正共同原因会与处理和结果同时关联。

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

      11.5.2 Adding a dummy confounder One way to test our models is to add dummy common-cause confounders. If a variable is not a confounder, it has no bearing on the true ATE, so we assume that our causal effect estimation workflow will be unaffected by these variables. In truth, additional variables might add statistical noise that throws off our estimator. The following listing attempts to

    参考来源

    - 位置:《Causal AI》第11章《构建因果推断工作流》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-11.xhtml) - 结论依据:摘录中关于随机协变量测试数值稳健性的定义、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:11.5.2 Adding a dummy confounder One way to test our models is to add dummy common-cause confounders. If a variable is not a confounder, it has no bearing on the true ATE, so we assume that our causal effect estimation workflow will be unaffected by these variables. In truth, additional variables might add statistical noise that throws off our estimator. The following listing attempts to