知识卡片
后验条件不保证每例更集中
事实信息可能让某次反事实分布更集中,但不是每个证据事件都使方差下降的定理。条件后验可选出高变异或多峰区域;应实际报告所求分布,不把比总体干预更窄当成正确性的必需标志。
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
原书出处 · 1 处
- · 查看原书依据
generally has much less spread than an interventional distribution representing the same hypothetical conditions. The counterfactual distribution essentially filters the interventional distribution down to cases where the conditions observed in the actual world are true. In this case, we have two height bell curves corresponding to two sexes. Those bell curves have a stronger overlap in the interventional distribution.
参考来源
- 位置:《Causal AI》第09章《通用反事实推断算法》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-9.xhtml)
- 结论依据:摘录中关于后验条件不保证每例更集中的定义、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:generally has much less spread than an interventional distribution representing the same hypothetical conditions. The counterfactual distribution essentially filters the interventional distribution down to cases where the conditions observed in the actual world are true. In this case, we have two height bell curves corresponding to two sexes. Those bell curves have a stronger overlap in the interventional distribution.