Fast Restricted Causal Inference
This work addresses a specific issue in causal inference for researchers, but it is incremental as it builds on an existing method.
The paper tackles the problem of hidden variables disturbing belief network recovery by introducing a new algorithm that accelerates the known CI algorithm, proving it does not produce spurious conditional independencies if statistical tests are reliable, unlike the FCI algorithm.
Hidden variables are well known sources of disturbance when recovering belief networks from data based only on measurable variables. Hence models assuming existence of hidden variables are under development. This paper presents a new algorithm "accelerating" the known CI algorithm of Spirtes, Glymour and Scheines {Spirtes:93}. We prove that this algorithm does not produces (conditional) independencies not present in the data if statistical independence test is reliable. This result is to be considered as non-trivial since e.g. the same claim fails to be true for FCI algorithm, another "accelerator" of CI, developed in {Spirtes:93}.