rob.fitConGraph {rggm}R Documentation

Fitting a concentration graph model based on a robustified method

Description

You can obtain robustified estimates of a Gaussian concentration graph model.

Usage

  rob.fitConGraph(amat, w.par, smpl.frm, it.limit = 200, tol = 1e-06)

Arguments

amat An adjacent matrix
w.par A non-negative real value, a robustness tuning parameter
smpl.frm A data frame, a sample data
it.limit a positive integer indicating the limit times of the iteration
tol A small positive number idicating the tolerance used in the convergence tests

Details

it.limit
When the case the number of the iteration is over it.limit, the fitting procedure will be turminated.

Value

mhat Estimates of the mean vector
Shat Estimates of the covariance matrix
w.vec Estimates of the weight vector
w.par The specified robustness tuning parameter
it The number of the iteration
tstat The test statistics for the overall goodness-of-fit
df The degree of freedom for the model
p.val The p-value of the model

Author(s)

Masashi Miyamura

References

Masashi Miyamura and Yutaka Kano, 2006, Robust Gaussian graphical modeling, Journal of Multivariate Analysis, 97, 7, 1525–1550

Examples

  require(mvtnorm)
  R <- matrix(c(1.0, 0.8, 0.0,
                0.8, 1.0, 0.0,
                0.0, 0.0, 1.0),
              3, 3)
  # A sample data including five outliers
  df <- as.data.frame(rbind(rmvnorm(95, rep(0, 3), R),
                            rmvnorm(5, rep(1, 3), R)))

  ug <- UG(~a*b*c)
  names(df) <- vertices(ug)

  rob.fitConGraph(ug, 0.3, df)

[Package rggm version 1.0.1 Index]