Determine support of multiple partial correlation/precision matrices

Share:

Description

A simple wrapper for sparsify which determines the support of a list of partial correlation/precision matrix by various methods and returns the sparsified matrices.

Usage

1

Arguments

Plist

A list of numeric precision matrices.

...

Arguments passed to sparsify.

Details

See sparsify for details.

Value

A list of the same length as Plist with the output from sparsify.

Author(s)

Anders Ellern Bilgrau, Wessel N. van Wierigen, Carel F.W. Peeters <cf.peeters@vumc.nl>

See Also

sparsify

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
ns <- c(10, 11)
Ylist <- createS(ns, p = 16, dataset = TRUE)
Slist <- lapply(Ylist, covML)
Tlist <- default.target.fused(Slist, ns)

# Obtain regularized precision under optimal penalty
opt <- optPenalty.fused.auto(Ylist, Tlist, cv.method = "aLOOCV",
                            maxit.ridgeP.fused = 1500)
# Use the optimal penalties
Plist <- ridgeP.fused(Slist, ns, lambda = opt$lambda, maxit = 1000)

# Determine support regularized (standardized) precision under optimal penalty
res <- sparsify.fused(Plist, threshold = "top", verbose = FALSE)
round(res[[1]]$sparsePrecision, 1)
round(res[[2]]$sparsePrecision, 1)

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.