spadd.presmth.Bspl.cv: Choose tuning parameters of desparsified lasso presmoothing...

In gregorkb/spaddinf: Fit two-step estimator for inference in sparse high-dimensional additive models

Description

Choose tuning parameters of desparsified lasso presmoothing estimator with cubic B-splines

Usage

spadd.presmth.Bspl.cv(X, Y, d.pre, n.lambda, n.eta, n.folds, plot = FALSE)

Arguments

X	the design matrix
Y	the response vector
d.pre	the number of intervals in which to divide the support of each covariate
n.lambda	the number of candidate lambda values
n.eta	the number of candidate eta values
n.folds	the number of crossvalidation folds

Value

a list with the chosen values of the tuning parameters

Examples

data <- data_gen(n = 200,q = 50,r = .9)

spadd.presmth.Bspl.cv.out <- spadd.presmth.Bspl.cv(X = data$X,
                                                   Y = data$Y,
                                                   d.pre = 20,
                                                   n.lambda = 25,
                                                   n.eta = 25,
                                                   n.folds = 5,
                                                   plot = TRUE)

spadd.presmth.Bspl.out <- spadd.presmth.Bspl(X = data$X,
                                             Y = data$Y,
                                             d.pre = 20,
                                             lambda = spadd.presmth.Bspl.cv.out$cv.lambda,
                                             eta = spadd.presmth.Bspl.cv.out$cv.eta,
                                             n.foi = 6)

plot_presmth_Bspl(spadd.presmth.Bspl.out)

gregorkb/spaddinf documentation built on July 23, 2021, 4:02 a.m.