A Constraint-Based Algorithm For Causal Discovery with Cycles, Latent Variables and Selection Bias

Causal processes in nature may contain cycles, and real datasets may violate\ncausal sufficiency as well as contain selection bias. No constraint-based\ncausal discovery algorithm can currently handle cycles, latent variables and\nselection bias (CLS) simultaneously. I therefore introduce an algorithm called\nCyclic Causal Inference (CCI) that makes sound inferences with a conditional\nindependence oracle under CLS, provided that we can represent the cyclic causal\nprocess as a non-recursive linear structural equation model with independent\nerrors. Empirical results show that CCI outperforms CCD in the cyclic case as\nwell as rivals FCI and RFCI in the acyclic case.\n