知识卡片
四种响应类型由成对结果定义
用(Y1,Y0)表示施加与不施加方案的二元结果,可分为受益者(1,0)、本来就成功者(1,1)、无论如何失败者(0,0)、被方案伤害者(0,1)。类型针对固定方案和结果定义,不是某个人永远不变的性格标签。
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
Segmenting users into persuadables, sure things, lost causes, and sleeping dogs In our churn example, we can assume there are two kinds of subscribers. For some subscribers, a promotion will influence their decision to churn. Others are non-responders, meaning people for whom the promotion will have no influence. We can break up the latter non-responders into two groups: Lost causes —People who will churn regardless
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:摘录中关于四种响应类型由成对结果定义的定义、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:Segmenting users into persuadables, sure things, lost causes, and sleeping dogs In our churn example, we can assume there are two kinds of subscribers. For some subscribers, a promotion will influence their decision to churn. Others are non-responders, meaning people for whom the promotion will have no influence. We can break up the latter non-responders into two groups: Lost causes —People who will churn regardless