cv.fitSTLS: cv.fitSTLS

In xinchoubiology/PALM: Solving Nonconvex and Nonsmooth Statistical Optimization Problems Using the PALM algorithm

Description

Cross-validation for Sparse Total Least Square model Using PALM algorithm

Usage

cv.fitSTLS(X, y, nfolds = 5, foldid = NULL, center = TRUE,
  lambda = NULL, nlambda = 100, lmin_ratio = 1e-04, alpha = 1,
  eps_abs = 1e-04, eps_rel = 1e-04, maxit = 1000L, warm_start = TRUE)

Arguments

X	The observation data matrix
y	The observation response vector
nfolds	Number of folds in cross-validation
foldid	Folds id in N-fold cross-validation
center	Y = Y - mean(Y) and X = X - mean(X)
lambda	The regularized lambda vector
nlambda	The Number of lambda
lmin_ratio	lambda_max / lambda_min
alpha	alpha parameter for elastic net penalty (0,1](Ridge Regression –> Lasso)
eps_abs	Absolutely epsilon for generated varaince vector
eps_rel	Relative epsilon for objective function value of generated vector
maxit	The maximal iteration of PALM algorithm
warm_start	Warm start for regularizer parameter tuning

Value

STLS_fit object

xinchoubiology/PALM documentation built on May 24, 2019, 9:56 a.m.