Shipley.test: Missing edge testing implied by a DAG with Shipley's...

In SEMgraph: Network Analysis and Causal Inference Through Structural Equation Modeling

Missing edge testing implied by a DAG with Shipley's basis-set

Description

Compute all the P-values of the d-separation tests implied by the missing edges of a given acyclic graph (DAG). The conditioning set Z is represented, in a DAG, by the union of the parent sets of X and Y (Shipley, 2000). The results of every test, in a DAG, is then combined using the Fisher’s statistic in an overall test of the fitted model C = -2*sum(log(P-value(k))), where C is distributed as a chi-squared variate with df = 2k, as suggested by Shipley (2000).

Usage

Shipley.test(
  graph,
  data,
  MCX2 = FALSE,
  cmax = Inf,
  limit = 100,
  verbose = TRUE,
  ...
)

Arguments

graph	A directed graph as an igraph object.
data	A data matrix with subjects as rows and variables as columns.
MCX2	If TRUE, a Monte Carlo P-value of the combined C test is enabled using the R code of Shipley extracted from <https://github.com/BillShipley/CauseAndCorrelation>.
cmax	Maximum number of parents set, C. This parameter can be used to perform only those tests where the number of conditioning variables does not exceed the given value. High-dimensional conditional independence tests can be very unreliable. By default, cmax = Inf.
limit	An integer value corresponding to the graph size (vcount) tolerance. Beyond this limit, multicore computation is enabled to reduce the computational burden. By default, limit = 100.
verbose	If TRUE, Shipley's test results will be showed to screen (default = TRUE).
...	Currently ignored.

Value

A list of three objects: (i) "dag": the DAG used to perform the Shipley test (ii) "dsep": the data.frame of all d-separation tests over missing edges in the DAG and (iii) "ctest": the overall Shipley's' P-value.

Author(s)

Mario Grassi [e-mail]

References

Shipley B (2000). A new inferential test for path models based on DAGs. Structural Equation Modeling, 7(2): 206-218. <https://doi.org/10.1207/S15328007SEM0702_4>

Examples

#\donttest{
# Nonparanormal(npn) transformation
als.npn <- transformData(alsData$exprs)$data

sem <- SEMrun(alsData$graph, als.npn)
C_test <- Shipley.test(sem$graph, als.npn, MCX2 = FALSE)
#MC_test <- Shipley.test(sem$graph, als.npn, MCX2 = TRUE)
#}

SEMgraph documentation built on Sept. 11, 2024, 8:36 p.m.