知识卡片
赋值函数应服务查询机制
选择线性、饱和曲线或程序分支,不应只因为某个库方便,而应表达目标推断所依赖的机制条件。模型不必逐行复制现实,但必须在关键查询上保留所需性质;拟合可接受不自动证明这些性质正确。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 结构因果模型
原书出处 · 1 处
- · 查看原书依据
Remember, you are modeling a ground-truth SCM. You are probably going to specify your assignment functions differently from those in the ground-truth SCM, but that’s fine. You don’t need your SCM to match the ground truth exactly; you just need your model to be right about the “how” assumptions it is relying on for your causal inferences.
参考来源
- 位置:《Causal AI》第06章《结构因果模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-6.xhtml)
- 结论依据:摘录中关于赋值函数应服务查询机制的定义、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:Remember, you are modeling a ground-truth SCM. You are probably going to specify your assignment functions differently from those in the ground-truth SCM, but that’s fine. You don’t need your SCM to match the ground truth exactly; you just need your model to be right about the “how” assumptions it is relying on for your causal inferences.