calcPredictorOK: Generate smoothing predictor given smoothing parameters

In blackbox: Black Box Optimization and Exploration of Parameter Space

Generate smoothing predictor given smoothing parameters

Description

Assuming that calcGCV has been first run to estimate smoothing parameter, this produces a “Kriging” predictor of the response.

Usage

calcPredictorOK(FONKgpointls, minKrigPtNbr = blackbox.getOption("minKrigPtNbr"),
                krigmax = NULL, topmode = FALSE, rawPlots = TRUE, cleanResu = "")

Arguments

FONKgpointls	Input data frame as produced by buildFONKgpointls
minKrigPtNbr	NULL or numeric. At least this many rows (if available) should be selected for Kriging. The default value depends on the number p of predictor variables and is 90, 159, 500, 1307, 3050, 6560 for p from 1 to 6 (beyond which it is strongly advised to use a non-default value).
krigmax	NULL or Numeric. For large data sets the selected points are not “Kriged” all together. Rather, overlapping blocks of rows are selected and are Kriged separately. This sets the size of the blocks. Default depends on the operating system (see source code).
topmode	Controls the way rows are selected. For development purposes, should not be modified
rawPlots	Boolean. Whether to plot one-dimensional “profiles” of the raw data.
cleanResu	A connection, or a character string naming a file for some nicely formated output. If "" (the default), print to the standard output connection.

Value

Returns invisibly a list with many undocumented elements. Thislist is also stored as a global option "fitobject".

blackbox documentation built on May 29, 2024, 1:15 a.m.