Accounting for hidden common causes when inferring cause and effect from\n observational data

2018/01/02 by David Heckerman, Heckerman, David · 1 citation
Computer Science · #Bayesian Modeling and Causal Inference

paper · pdf · doi:10.48550/arxiv.1801.00727

Abstract

Identifying causal relationships from observation data is difficult, in large\npart, due to the presence of hidden common causes. In some cases, where just\nthe right patterns of conditional independence and dependence lie in the\ndata---for example, Y-structures---it is possible to identify cause and effect.\nIn other cases, the analyst deliberately makes an uncertain assumption that\nhidden common causes are absent, and infers putative causal relationships to be\ntested in a randomized trial. Here, we consider a third approach, where there\nare sufficient clues in the data such that hidden common causes can be\ninferred.\n

Cited by

• The Blessings of Multiple Causes

Related

• Measuring Latent Causal Structure
• Leveraging directed causal discovery to detect latent common causes
• Thinking Clearly About Correlations and Causation: Graphical Causal Models for Observational Data
• On Multi-Cause Causal Inference with Unobserved Confounding: Counterexamples, Impossibility, and Alternatives
• Detection of Unobserved Common Causes based on NML Code in Discrete, Mixed, and Continuous Variables