Supplementary material for Keleș and Chun (2008)
TEST, Volume 17, Issue 1
Comments on: "Augmenting the bootstrap to analyze high dimensional genomic data" by Tyekucheva and Chiaromonte.
Connections to the ridge regularized covariance estimator with bagging
Code: function.R (to be loaded into the R session)
run.R (actual script to run simulations)
Mean Relatived Squared Errors (RSEs) for various values of the smoothing (tuning) parameter [0.001, 1.2].
A total of 500 simulations is generated for each scenario.
Panels A, B, and C correspond to sample sizes n = 10p, n = 1.1p, and n = 0.5p, respectively.