with_graph: Evaluate Causal Graph Discovery Algorithm in Multiple Imputed...

In bips-hb/micd: Multiple Imputation in Causal Graph Discovery

Evaluate Causal Graph Discovery Algorithm in Multiple Imputed Data sets

Description

Evaluate Causal Graph Discovery Algorithm in Multiple Imputed Data sets

Usage

with_graph(data, algo = c("pc", "fci", "fciPlus", "ges"), args, score = FALSE)

Arguments

data	An object of type mids, which stands for 'multiply imputed data set', typically created by a call to function mice()
algo	An algorithm for causal discovery from the package 'pcalg' (see details).
args	Additional arguments passed to the algo. Must be a string vector starting with comma, i.e. ", ..."
score	Logical indicating whether a score-based or a constrained-based algorithm is applied.

Value

A list object of S3 class mice::mira-class.

Examples

data(windspeed)
dat <- as.matrix(windspeed)

## delete some observations
set.seed(123)
dat[sample(1:length(dat), 260)] <- NA

## Impute missing values under normal model
imp <- mice(dat, method = "norm", printFlag = FALSE)
mylabels <- names(imp$imp)
out.fci <- with_graph(data = imp, 
                      algo = "fciPlus", 
                      args = ", indepTest = gaussCItest, verbose = FALSE,
                      labels = mylabels, alpha = 0.01")
                          
 out.ges <- with_graph(data = imp, algo = "ges", arg = NULL, score = TRUE)
 
if (requireNamespace("Rgraphviz", quietly = TRUE)){
 oldpar <- par(mfrow = c(1,2)) 
   plot(out.fci$res[[1]])
   plot(out.ges$res[[1]]$essgraph)
 par(oldpar)
 }
                                     

bips-hb/micd documentation built on Feb. 22, 2023, 7:59 p.m.