Deriving Complete Constraints in Hidden Variable Models
Hidden variable graphical models can sometimes imply constraints on the observable distribution that are more complex than simple conditional independence relations. These observable constraints can allow for falsification of assumptions of the model that would otherwise be untestable due to the unobserved variables and can be used to constrain estimation procedures to improve efficiency. Knowing the complete set of observable constraints is thus ideal, but this can be difficult to determine in many settings. In models with categorical observed variables and a joint distribution that is completely characterized by linear relations to the unobservable response function variables, we develop a systematic method for deriving the complete set of observable constraints. This package implements that method and provides some examples in the vignette.
This is the companion R package to the following paper: https://arxiv.org/pdf/2601.11242.
This package is in early development. The documentation is incomplete, and the API is subject to change.
Our package meraconstraints depends on caugi to create the graphical models and do some operations. caugi is very fast, user friendly, and supports mixed graphs. Read more about caugi here: https://caugi.org or on GitHub here https://github.com/frederikfabriciusbjerre/caugi/.
Installation
For now it is only available on github.
remotes::install_github("sachsmc/meraconstraints")