BriefGPT.xyz
Jul, 2022
选择偏差下的边界反事实
Bounding Counterfactuals under Selection Bias
HTML
PDF
Marco Zaffalon, Alessandro Antonucci, Rafael Cabañas, David Huber, Dario Azzimonti
TL;DR
本研究提出一种算法来解决因果分析中的选择偏差,并证明了尽管存在选择偏差,可用数据的似然函数是单峰的。该算法可以用于解决可识别和不可识别查询,并通过因果期望最大化方案计算可识别情况下的因果查询值,否则计算上下界。实验表明该方法是实际可行的,并提供了理论收敛特性。
Abstract
causal analysis
may be affected by
selection bias
, which is defined as the systematic exclusion of data from a certain subpopulation. Previous work in this area focused on the derivation of
→