知识卡片
条件化保留原来的生成选择
观察到E=high是在原DGP中选出自然高参与者;do(E=high)把目标人群的E改为high,保留其他机制。若G同时影响E和I,两者中的G人群构成通常不同,这正是条件分布与干预分布可不同的原因。
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
- · 干预与因果效应
原书出处 · 1 处
- · 查看原书依据
P(I|E=“high”) is the probability distribution of In-Game Purchases given players with “high” engagement, and P(I|do(E= “high”)) is the probability distribution of In-Game Purchases given an intervention that sets a player’s engagement level to “high.” The second column of table 7.11 illustrates extensions of this notation to joint distributions, multiple interventions, and mixing interventions with observations.
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:摘录中关于条件化保留原来的生成选择的定义、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:P(I|E=“high”) is the probability distribution of In-Game Purchases given players with “high” engagement, and P(I|do(E= “high”)) is the probability distribution of In-Game Purchases given an intervention that sets a player’s engagement level to “high.” The second column of table 7.11 illustrates extensions of this notation to joint distributions, multiple interventions, and mixing interventions with observations.