WH (version 1.1.2)

WH_1d_perf: Whittaker-Henderson Smoothing (Maximum Likelihood, Generalized Fellner-Schall update)

Description

Whittaker-Henderson Smoothing (Maximum Likelihood, Generalized Fellner-Schall update)

Usage

WH_1d_perf(
  d,
  ec,
  y,
  wt,
  q = 2,
  p,
  criterion = "REML",
  lambda = 1000,
  reg = FALSE,
  verbose = FALSE,
  accu_crit = 1e-12,
  accu_dev = 1e-12
)

Value

An object of class "WH_1d" i.e. a list containing model fit, variance, residuals and degrees of freedom as well as diagnosis to asses the quality of the fit.

Arguments

d

Vector of observed events

ec

Vector of central exposure

y

Vector of observations

wt

Optional vector of weights

q

Order of penalization. Polynoms of degrees q - 1 are considered smooth and are therefore unpenalized

p

The number of eigenvectors to keep

criterion

Criterion used to choose the smoothing parameter. One of "GCV" (default), "AIC" or "BIC".

lambda

Initial smoothing parameter

reg

Should the regression framework be used ? Boolean. If TRUE, will stop after the first iteration.

verbose

Should information about the optimization progress be displayed

accu_crit

Tolerance for the convergence of the outer optimization procedure

accu_dev

Tolerance for the convergence of the optimization procedure