Compute initial estimates for the EN S-estimator using the EN-PY procedure.
enpy_initial_estimates( x, y, alpha, lambda, bdp = 0.25, cc, intercept = TRUE, penalty_loadings, enpy_opts = enpy_options(), mscale_opts = mscale_algorithm_options(), eps = 1e-06, sparse = FALSE, ncores = 1L )
vector of response values of length
elastic net penalty mixing parameter with \(0 \le \alpha \le 1\).
a vector of positive values of penalization levels.
desired breakdown point of the estimator, between 0 and 0.5. The actual breakdown point may be slightly larger/smaller to avoid instabilities of the S-loss.
cutoff value for the bisquare rho function. By default, chosen to yield a consistent estimate for the Normal distribution.
include an intercept in the model.
a vector of positive penalty loadings (a.k.a. weights) for different
penalization of each coefficient. Only allowed for
options for the EN-PY algorithm, created with the
options for the M-scale estimation. See
use sparse coefficient vectors.
number of CPU cores to use in parallel. By default, only one CPU core is used. Not supported on all platforms, in which case a warning is given.
If these manually computed initial estimates are intended as starting points for
they are by default shared for all penalization levels.
To restrict the use of the initial estimates to the penalty level they were computed for, use
as_starting_point(..., specific = TRUE). See
as_starting_point() for details.
Cohen Freue, G.V.; Kepplinger, D.; Salibián-Barrera, M.; Smucler, E. Robust elastic net estimators for variable selection and identification of proteomic biomarkers. Ann. Appl. Stat. 13 (2019), no. 4, 2065--2090 doi: 10.1214/19-AOAS1269