Semiparametric Inference For Causal Effects In Graphical Models With Hidden Variables

被引:0
|
作者
Bhattacharya, Rohit [1 ]
Nabi, Razieh [2 ]
Shpitser, Ilya [3 ]
机构
[1] Williams Coll, Dept Comp Sci, Williamstown, MA 01267 USA
[2] Emory Univ, Dept Biostat & Bioinformat, Atlanta, GA 30322 USA
[3] Johns Hopkins Univ, Dept Comp Sci, Baltimore, MD 21218 USA
关键词
Unmeasured confounding; doubly robust estimation; nonparametric satura-tion; efficient influence function; DOUBLY ROBUST ESTIMATION; EFFICIENT; DIAGRAMS;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Identification theory for causal effects in causal models associated with hidden variable directed acyclic graphs (DAGs) is well studied. However, the corresponding algorithms are underused due to the complexity of estimating the identifying functionals they output. In this work, we bridge the gap between identification and estimation of population-level causal effects involving a single treatment and a single outcome. We derive influence function based estimators that exhibit double robustness for the identified effects in a large class of hidden variable DAGs where the treatment satisfies a simple graphical criterion; this class includes models yielding the adjustment and front-door functionals as special cases. We also provide necessary and sufficient conditions under which the statistical model of a hidden variable DAG is nonparametrically saturated and implies no equality constraints on the observed data distribution. Further, we derive an important class of hidden variable DAGs that imply observed data distributions observationally equivalent (up to equality constraints) to fully observed DAGs. In these classes of DAGs, we derive estimators that achieve the semiparametric efficiency bounds for the target of interest where the treatment satisfies our graphical criterion. Finally, we provide a sound and complete identification algorithm that directly yields a weight based estimation strategy for any identifiable effect in hidden variable causal models.
引用
收藏
页码:1 / 76
页数:76
相关论文
共 50 条
  • [1] Semiparametric Inference for Causal Effects in Graphical Models with Hidden Variables
    Bhattacharya, Rohit
    Nabi, Razieh
    Shpitser, Ilya
    [J]. Journal of Machine Learning Research, 2022, 23
  • [2] Causal graphical models with latent variables: Learning and inference
    Meganck, Stijn
    Leray, Philippe
    Manderick, Bernard
    [J]. SYMBOLIC AND QUANTITATIVE APPROACHES TO REASONING WITH UNCERTAINTY, PROCEEDINGS, 2007, 4724 : 5 - +
  • [3] Efficient adjustment sets in causal graphical models with hidden variables
    Smucler, E.
    Sapienza, F.
    Rotnitzky, A.
    [J]. BIOMETRIKA, 2022, 109 (01) : 49 - 65
  • [4] Necessary and sufficient graphical conditions for optimal adjustment sets in causal graphical models with hidden variables
    Runge, Jakob
    [J]. ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 34 (NEURIPS 2021), 2021, 34
  • [5] A Bayesian Semiparametric Approach to Intermediate Variables in Causal Inference
    Schwartz, Scott L.
    Li, Fan
    Mealli, Fabrizia
    [J]. JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 2011, 106 (496) : 1331 - 1344
  • [6] Graphical models, causal inference, and econometric models
    Spirtes, Peter
    [J]. JOURNAL OF ECONOMIC METHODOLOGY, 2005, 12 (01) : 3 - 34
  • [7] Graphical Causal Models for Survey Inference
    Schuessler, Julian
    Selb, Peter
    [J]. SOCIOLOGICAL METHODS & RESEARCH, 2023,
  • [8] Bayesian Causal Inference in Probit Graphical Models
    Castelletti, Federico
    Consonni, Guido
    [J]. BAYESIAN ANALYSIS, 2021, 16 (04): : 1113 - 1137
  • [9] Identification in Causal Models With Hidden Variables
    Shpitser, Ilya
    [J]. JOURNAL OF THE SFDS, 2020, 161 (01): : 91 - 119
  • [10] COPULA GAUSSIAN GRAPHICAL MODELS WITH HIDDEN VARIABLES
    Yu, Hang
    Dauwels, Justin
    Wang, Xueou
    [J]. 2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2012, : 2177 - 2180