Apply a kernel implied in an emc object

This function extracts the appropriate trend, and applies its implied kernel using either the pure R implementation or the Rcpp version. When mode = "compare", the function checks whether the two implementations produce identical output.

Usage

apply_kernel(
  kernel_pars,
  emc,
  subject = 1,
  input_pars = NULL,
  trend_n = 1,
  mode = "Rcpp"
)

Arguments

kernel_pars

A named vector of kernel parameters on the natural scale. Use NULL for kernels that do not require parameters.

emc

An emc.

subject

Subject index for which to apply the kernel. Defaults to 1.

input_pars

Optional parameter matrix containing externally supplied parameter values (e.g., trend parameters). Only needed for custom kernels.

trend_n

Integer specifying which trend to apply when multiple trends exist in the model. Defaults to 1. A warning is issued if the model contains more than one trend.

mode

Character string specifying which implementation to use:

"R"

Use the pure R implementation.

"Rcpp"

Use the Rcpp implementation (default).

Value

Returns a kernel matrix produced by the corresponding implementation.

Details

Applies the trend-specific kernel associated with an emc model to the a subject's data and returns the resulting kernel matrix.