quic.graphical_model {stabs}R Documentation

Stability selection fit function for sparse inverse covariance using QUIC

Description

Stability selection fit function for sparse inverse covariance using package QUIC.

Usage

quic.graphical_model(x, y, q, ...)
getLamPath(max, min, len, log = FALSE)

Arguments

x

data matrix

y

data matrix

q

number of variables

...

additional arguments passed to the underlying fitting function.

max

maximum value for regularization (lambda)

min

min value for lambda

len

length of path

log

log spacing

Details

This is a wrapper for QUIC to be used in stability selection. Pass it as the fit function to stabsel. If you supply your own graphical model fitter, please set the class to "graphical_model".

getLamPath can be used to create a regularization path.

Value

A named list with elements

selected

logical. A vector that indicates which variable was selected.

path

logical. A matrix that indicates which variable was selected in which step. Each row represents one variable, the columns represent the steps.

Author(s)

Code contributed by Richard Beare.

See Also

stabsel for stability selection itself, and fitfun for other fitting functions.

Examples

### Do not test the following code per default on CRAN as it takes some time to run:
if (require("huge") && require("QUIC")) {
  set.seed(10010)
  dat.hubs <- huge.generator(n = 1000, d = 40, graph = "hub")
  s.hubs <- stabsel(x = dat.hubs$data, fitfun = quic.graphical_model, 
                    cutoff = 0.75, PFER = 10)
}


[Package stabs version 0.6-3 Index]