PC (rank) search

Source: R/causal_inference_helpers.R

Runs PC (rank) algorithm given a dataset dat.

pc_rank_search(dat, alpha = 0.05)

Arguments

dat

Numeric matrix. Dataset matrix with n rows (observations) and p columns (variables).

alpha

Numeric — between 0 and 1. The significance level for the individual conditional independence tests. By default it is set to 0.05.

Value

Square binary matrix (or NA in case of error). The CPDAG estimated from the data.

Details

The PC (rank) algorithm has been proposed by Harris, N., and Drton, M., http://jmlr.csail.mit.edu/papers/volume14/harris13a/harris13a.pdf

The function pcalg:pc() is called with the following arguments:

  • suffStat = list(C = 2 * sin(cor(dat, method = "spearman") * pi/6), n = nrow(dat)),

  • indepTest = gaussCItest,

  • u2pd = "retry" — this ensures that the produced CPDAG is extendible to a DAG,

  • skel.method = "stable".