kbackfit {gplm}

Backfitting for an additive model using kernel regression

Description

Implements kernel-based backfitting in an additive model, optional with a partial linear term.

Usage

kbackfit(t, y, h, x = NULL, grid = NULL, weights.conv = 1,
           offset = 0, method = "generic",
           max.iter = 50, eps.conv = 1e-04, m.start = NULL,
           kernel = "biweight")

Arguments

y

n x 1 vector, responses

t

n x q matrix, data for nonparametric part

h

scalar or 1 x q, bandwidth(s)

x

optional, n x p matrix, data for linear part

grid

m x q matrix, where to calculate the nonparametric function (default = t)

weights.conv

weights for convergence criterion

offset

offset

method

one of "generic", "linit" or "modified"

max.iter

maximal number of iterations

eps.conv

convergence criterion

m.start

n x q matrix, start values for m

kernel

text string, see kernel.function

Value

List with components:

c

constant

b

p x 1 vector, linear coefficients

m

n x q matrix, nonparametric marginal function estimates

m.grid

m x q matrix, nonparametric marginal function estimates on grid

rss

residual sum of squares

Author(s)

Marlene Mueller

See Also

kernel.function, kreg