cFuSIM_index: This function computes the coeffcients of the index function

In YunlongNie/cFuSIM: R package for Functional Single Index Model with compact support

Description

This function computes the coeffcients of the index function

Usage

cFuSIM_index(y, xfd, spline_basis, threshold = 1e-05, maxit = 150,
  lambda = 10000)

Arguments

y	a vector of response.
xfd	a fda object contains the functional predictor
spline_basis	the B-spline basis for the index function
lambda	a positive number, the sparsity parameter

Examples

## Not run: 
library(fda)
library(cFuSIM)
data(bike_cFuSIM)
timepts = bike$timepts
norder=4 ## cubic B-spline
nbasis=norder+length(timepts)-2; 
spline_basis=create.bspline.basis(rangeval=c(1,24),nbasis#'norder,timepts)
wull = bike$temp
xfds=  Data2fd(y=wull%>%t, argvals=bike$timepts)
y = bike$y
train_sample = 1:length(y)
y = y[train_sample]
xfd = xfds[train_sample]
res_c = cFuSIM_index(y, xfd, spline_basis)
beta_fd = fd(res_c$coefBeta, res_c$basisBeta)
plot(beta_fd,ylab="index function", xlab='time')
fdagg(beta_fd)
score_fit = (res_c$score_fit)
pred_y = localpoly.reg(score_fit, y, degree.pol = 1, kernel.type = "gaussian",bandwidth = "CV",deriv=0,points#'score_fit)
plot(x=score_fit, y=y)
lines(pred_y$predicted[order(score_fit)],x=score_fit[order#'score_fit)],col=4)

## End(Not run)

YunlongNie/cFuSIM documentation built on Feb. 8, 2021, 5:51 p.m.