Nonparametric Causal Structure Learning in High Dimensions

被引:1
作者
Chakraborty, Shubhadeep [1 ]
Shojaie, Ali [1 ]
机构
[1] Univ Washington, Dept Biostat, Seattle, WA 98195 USA
基金
美国国家科学基金会; 美国国家卫生研究院;
关键词
causal structure learning; consistency; FCI algorithm; high dimensionality; nonparametric testing; PC algorithm; DIRECTED ACYCLIC GRAPHS; DISTANCE CORRELATION; SELECTION; LATENT;
D O I
10.3390/e24030351
中图分类号
O4 [物理学];
学科分类号
0702 ;
摘要
The PC and FCI algorithms are popular constraint-based methods for learning the structure of directed acyclic graphs (DAGs) in the absence and presence of latent and selection variables, respectively. These algorithms (and their order-independent variants, PC-stable and FCI-stable) have been shown to be consistent for learning sparse high-dimensional DAGs based on partial correlations. However, inferring conditional independences from partial correlations is valid if the data are jointly Gaussian or generated from a linear structural equation model-an assumption that may be violated in many applications. To broaden the scope of high-dimensional causal structure learning, we propose nonparametric variants of the PC-stable and FCI-stable algorithms that employ the conditional distance covariance (CdCov) to test for conditional independence relationships. As the key theoretical contribution, we prove that the high-dimensional consistency of the PC-stable and FCI-stable algorithms carry over to general distributions over DAGs when we implement CdCov-based nonparametric tests for conditional independence. Numerical studies demonstrate that our proposed algorithms perform nearly as good as the PC-stable and FCI-stable for Gaussian distributions, and offer advantages in non-Gaussian graphical models.
引用
收藏
页数:23
相关论文
共 50 条
[41]   Learning causal graphs via nonlinear sufficient dimension reduction [J].
Solea, Eftychia ;
Li, Bing ;
Kim, Kyongwon .
JOURNAL OF MACHINE LEARNING RESEARCH, 2025, 26 :1-46
[42]   Load-Balanced Parallel Constraint-Based Causal Structure Learning on Multi-Core Systems for High-Dimensional Data [J].
Schmidt, Christopher ;
Huegle, Johannes ;
Bode, Philipp ;
Uacker, Matthias .
2019 ACM SIGKDD WORKSHOP ON CAUSAL DISCOVERY, VOL 104, 2019, 104 :59-77
[43]   Progressive Skeleton Learning for Effective Local-to-Global Causal Structure Learning [J].
Guo, Xianjie ;
Yu, Kui ;
Liu, Lin ;
Li, Jiuyong ;
Liang, Jiye ;
Cao, Fuyuan ;
Wu, Xindong .
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2024, 36 (12) :9065-9079
[44]   Learning causal structure from mixed data with missing values using Gaussian copula models [J].
Cui, Ruifei ;
Groot, Perry ;
Heskes, Tom .
STATISTICS AND COMPUTING, 2019, 29 (02) :311-333
[45]   Clinical causal analysis via iterative active structure learning [J].
Tao, Zhenchao ;
Chi, Meiyan ;
Chen, Lyuzhou ;
Ban, Taiyu ;
Tu, Qiang ;
Gao, Fei ;
Wang, Wei .
MEMETIC COMPUTING, 2025, 17 (01)
[46]   Towards Cross-Modal Causal Structure and Representation Learning [J].
Mao, Haiyi ;
Liu, Hongfu ;
Dou, Jason Xiaotian ;
Benos, Panayiotis V. .
MACHINE LEARNING FOR HEALTH, VOL 193, 2022, 193 :120-140
[47]   Conditions and Assumptions for Constraint-based Causal Structure Learning [J].
Sadeghi, Kayvan ;
Soo, Terry .
JOURNAL OF MACHINE LEARNING RESEARCH, 2022, 23
[48]   Bootstrap-Based Layerwise Refining for Causal Structure Learning [J].
Xiang G. ;
Wang H. ;
Yu K. ;
Guo X. ;
Cao F. ;
Song Y. .
IEEE Transactions on Artificial Intelligence, 2024, 5 (06) :2708-2722
[49]   Causal Learning via Manifold Regularization [J].
Hill, Steven M. ;
Oates, Chris J. ;
Blythe, Duncan A. ;
Mukherjee, Sach .
JOURNAL OF MACHINE LEARNING RESEARCH, 2019, 20
[50]   Efficient Learning of Minimax Risk Classifiers in High Dimensions [J].
Bondugula, Kartheek ;
Mazuelas, Santiago ;
Perez, Aritz .
UNCERTAINTY IN ARTIFICIAL INTELLIGENCE, 2023, 216 :206-215