empirical_knots: Estimate knots for reduced rank spatial process

In jelsema/RRSM: Methods for reduced rank spatial models

Description

For a given bandwith, estimate knots for a reduced rank spatial models using the centroids of interval sets, iterating through all thresholds.

Usage

empirical_knots(coords, R, delta1, delta2 = delta1, slices, augment = FALSE,
  knots_grid = NULL, ...)

Arguments

coords	the observed locations.
R	the data to be used in estimating knots.
delta1	the first bandwidth, used to determine the knots for a given threshold interval set, passed to threshold_knots.
delta2	the second bandidth, used to detect if newly estimated knots are too close to existing knots.
slices	the threshold cut-off values to define interval sets.
augment	whether to augment the estimated set of knots with a regular grid of knots (fills in sparse areas). See Notes.
knots_grid	(optional) the grid of knots to use if augment=TRUE.
...	space for additional arguments.

Value

A matrix of locations.

Note

The argument augment was considered in the course of development of the Threshold Knot Selection (TKS) algorithm, but was not pursued with much depth. The properties of TKS when augmenting with a grid have not been studied.

See Also

TKS

jelsema/RRSM documentation built on May 19, 2019, 4:02 a.m.