知识卡片
启用停用概率使用不同条件
PD=P(Y0=0|Y=1),PE=P(Y1=1|Y=0),分别在结果已发生或未发生的人群中问移除或施加原因的替代结果。它们没有再固定实际X,因此与PN、PS的目标人群不同,不能直接互换。
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
- · 反事实与平行世界
原书出处 · 1 处
- · 查看原书依据
The probabilities of disablement (PD) and enablement (PE) are similar to PN and PS, except they do not condition on the cause X. PD is the query P (YX= 0= 0|Y = 1), meaning “For cases where Y happened, if X had not happened would Y not have happened?” For the churn problem, PD asks the question “What is the overall chance of churn if
参考来源
- 位置:《Causal AI》第08章《反事实与平行世界》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-8.xhtml)
- 结论依据:摘录中关于启用停用概率使用不同条件的定义、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:The probabilities of disablement (PD) and enablement (PE) are similar to PN and PS, except they do not condition on the cause X. PD is the query P (YX= 0= 0|Y = 1), meaning “For cases where Y happened, if X had not happened would Y not have happened?” For the churn problem, PD asks the question “What is the overall chance of churn if