enhance: Enhance penalty value

In hero: Spatio-Temporal (Hero) Sandwich Smoother

Enhance penalty value

Description

enhance enhances the sandwich smoother by choosing the optimal penalty value that minimizes the GCV statistic. The optimx function is used to do the optimization.

Usage

enhance(
  obj,
  par = rep(0, length(obj$n)),
  lower = rep(-20, length(par)),
  upper = rep(20, length(par)),
  method = "L-BFGS-B",
  control = list(),
  prepare = TRUE,
  loggcv = FALSE,
  ...
)

Arguments

obj	A prepared_* object from a prepare function.
par	a vector of initial values for the parameters for which optimal values are to be found. Names on the elements of this vector are preserved and used in the results data frame.
lower, upper	Bounds on the variables for methods such as "L-BFGS-B" that can handle box (or bounds) constraints.
method	The method to be used for optimization. The default is L-BFGS-B, which allows for constraints on the parameters to optimize. See optimx for all available methods.
control	A list of control parameters. See ‘Details’.
prepare	A logical value. The default is TRUE, indicating that a prepared_data object should be returned. If FALSE, then the results of the call to the optimx function is returned.
loggcv	A logical value indicating whether the log of the GCV statistic should be used. Useful for very large data sets. Default is TRUE.
...	Additional arguments to pass to to the optimx function.

Details

Internally, the loglambda2gcv is used as the objective function for the optimx function. Many different optimization methods are available. The default is L-BFGS-B, which allows for constraints on the parameters to optimize. Another excellent choice is the nlminb algorithm, which also allows for parameter constraints.

Value

By default, a prepared_data object with the optimal loglambda values that minimize the GCV, along with an additional component, results, that contains the optimization results.

Author(s)

Joshua French

Examples

# create b-splines
x1 = bspline(nbasis = 10)
x2 = bspline(nbasis = 12)

# observed data locations
evalarg1 = seq(0, 1, len = 60)
evalarg2 = seq(0, 1, len = 80)

# construct "true" data
mu = matrix(0, nrow = 60, ncol = 80)
for(i in seq_len(60)) {
   for(j in seq_len(80)) {
      mu[i, j] =  sin(2*pi*(evalarg1[i]-.5)^3)*cos(4*pi*evalarg2[j])
   }
}
# construct noisy data
data = mu + rnorm(60 * 80)

obj = prepare(data, list(evalarg1, evalarg2), list(x1, x2))
enhance(obj)

hero documentation built on July 26, 2023, 5:11 p.m.