Prior and posterior checking of implicit causal assumptions

Prior and posterior checking of implicit causal assumptions
复制标题

隐含因果假设的事前和事后检查

DOI:
10.1111/biom.13886
复制
发表时间:
2023
期刊:
影响因子:
1.9
通讯作者:
Linero, Antonio R.
Linero, Antonio R.
中科院分区:
数学3区
文献类型:
--
作者:
Linero, Antonio R.

文献摘要

参考文献

相似文献

因果推理从业者越来越多地采用机器学习技术,目的是对因果效应进行有原则的不确定性量化,同时最大限度地减少模型错误指定的风险。贝叶斯非参数方法也因其灵活性和提供自然不确定性量化的承诺而引起了人们的关注。然而,高维或非参数空间上的先验通常会无意中编码与因果推理中的实质性知识不一致的先验信息,具体来说,高维贝叶斯模型工作所需的正则化可能间接意味着混杂的程度可以忽略不计。在本文中,我们解释了这个问题并提供了工具来(i)验证先验分布没有编码远离混杂模型的归纳偏差,以及(ii)验证后验分布是否包含足够的信息来克服这个问题(如果存在)。我们对高维概率岭回归模型的模拟数据进行了概念验证,并说明了应用于大型医疗支出调查的贝叶斯非参数决策树集成。
Causal inference practitioners have increasingly adopted machine learning techniques with the aim of producing principled uncertainty quantification for causal effects while minimizing the risk of model misspecification. Bayesian nonparametric approaches have attracted attention as well, both for their flexibility and their promise of providing natural uncertainty quantification. Priors on high-dimensional or nonparametric spaces, however, can often unintentionally encode prior information that is at odds with substantive knowledge in causal inference—specifically, the regularization required for high-dimensional Bayesian models to work can indirectly imply that the magnitude of the confounding is negligible. In this paper, we explain this problem and provide tools for (i) verifying that the prior distribution does not encode an inductive bias away from confounded models and (ii) verifying that the posterior distribution contains sufficient information to overcome this issue if it exists. We provide a proof-of-concept on simulated data from a high-dimensional probit-ridge regression model, and illustrate on a Bayesian nonparametric decision tree ensemble applied to a large medical expenditure survey.
DOI: 10.1111/biom.13499
发表时间: 2021-06-06
期刊: BIOMETRICS
影响因子: 1.9
作者:
Linero, Antonio R.
通讯作者: Linero, Antonio R.
在非参数和高维模型中,贝叶斯可忽略性是一个信息丰富的先验
DOI: 10.1080/01621459.2023.2278202
发表时间: 2023
影响因子: 3.7
作者:
Linero, Antonio R.
通讯作者: Linero, Antonio R.
贝叶斯因果推理:批判性评论
DOI: 10.1098/rsta.2022.0153
发表时间: 2023
期刊: Physical and Engineering Sciences
影响因子: --
作者:
Li, Fan;Ding, Peng;Mealli, Fabrizia
通讯作者: Mealli, Fabrizia
DOI: 10.1038/s41586-019-1466-y
发表时间: 2019-09-19
期刊: NATURE
影响因子: 64.8
作者:
Yeager, David S.;Hanselman, Paul;Dweck, Carol S.
通讯作者: Dweck, Carol S.
DOI: 10.1145/3241036
发表时间: 2019-03-01
影响因子: 22.7
作者:
Pearl, Judea
通讯作者: Pearl, Judea