edgewiseTSID: Determines which edges in a mixed graph are edgewiseID+TS...
In Lucaweihs/SEMID: Identifiability of Linear Structural Equation Models

edgewiseTSID

R Documentation

Determines which edges in a mixed graph are edgewiseID+TS identifiable

Description

Uses the edgewise+TS identification criterion of Weihs, Robeva, Robinson, et al. (2017) to determine which edges in a mixed graph are generically identifiable. In particular this algorithm iterates between the half-trek, edgewise, and trek-separation identification algorithms in an attempt to identify as many edges as possible, this may be very slow.

Usage

edgewiseTSID(
  mixedGraph,
  tianDecompose = T,
  subsetSizeControl = 3,
  maxSubsetSize = 3
)

Arguments

`mixedGraph`	a `MixedGraph` object representing the L-SEM.
`tianDecompose`	TRUE or FALSE determining whether or not the Tian decomposition should be used before running the current generic identification algorithm. In general letting this be TRUE will make the algorithm faster and more powerful.
`subsetSizeControl`	a positive integer (Inf allowed) which controls the size of edgesets searched in the edgewiseID algorithm. Suppose, for example, this has value 3. Then if a node i has n parents, this will restrict the algorithm to only look at subsets of the parents of size 1,2,3 and n-2, n-1, n. Making this parameter smaller means the algorithm will be faster but less exhaustive (and hence less powerful).
`maxSubsetSize`	a positive integer which controls the maximum subset size considered in the trek-separation identification algorithm. Making this parameter smaller means the algorithm will be faster but less exhaustive (and hence less powerful).