minimize: Minimization rutine

In msalibian/RobustFPLM: Robust estimators for Functional Partial Linear Models

Description

This function is used internally by FPLMBsplines(), not meant for 'public' use.

Usage

minimize(
  yy,
  xx_coef,
  uu,
  spl_kn,
  freq,
  fLoss,
  norder,
  rob.control = lmrob.control(trace.level = 0, nResample = 5000, tuning.psi = 3.443689,
    subsampling = "simple", rel.tol = 1e-05, refine.tol = 1e-05, k.max = 2000, maxit.scale
    = 2000, max.it = 2000)
)

Arguments

yy	response.
xx_coef	covariates coefficients.
uu	values for thenon parametric term.
spl_kn	number of splines for the non parametric term.
freq	number of splines for the functional term.
fLoss	string specifying the loss function. 'ls' for least squares, 'huang' for Huber, 'lmbrob' for MM-estimator.
norder	spline order (n=4 for cubic splines)
rob.control	optional parameters if fLoss == 'lmbrob'.

Value

A vector with components:

• spl estimate for the non parametric term,

• slope estimate for the functional slope,

• value the minimum value of the loss

• scale the scale estimate.

msalibian/RobustFPLM documentation built on July 4, 2020, 5:46 p.m.