sampling {RGeostats}R Documentation

Derive a model-based sampling on a Data set

Description

Derive a model-based sampling on a Data set

Usage

sampling(db, model, beta = 0, mean = NA, nsize1 = 3, nsize2 = 10,
         method1 = 1, method2 = 1, ranks1 = NA, ranks2 = NA, verbose = FALSE)

Arguments

db

The db-class which contains the samples for which a sampling must be calculated

model

The model-class which defines the spatial characteristics of a variable defined on the Db. This model will impact the samples selected in the sampling. For more information check check.model.

beta

This parameter corresponds to the percentage of samples (although selected in a preliminary sampling) which must be ignored in order to reduce the number of retained pivots. This will reduce the computational time for a posterior kriging activity.

mean

The sampling is performed in (strict) stationary condition (Simple Kriging method). Then the user must provide the theoretical mean.

nsize1

Number of exact pivots to be retained in the sampling.

nsize2

Number of pivots (driven by ACP) to be retained in the sampling.

method1

Selection criterion for selecting the exact pivots (in the iterative selection algorithm):

  • 1 Local criterion: The next pivot is the sample which receives the largest estimation error (comparing the estimated and the true values).

  • 2 Global criterion: The next pivot is the sample such that, if it were included in the basis of pivots, it would lead to the smallest squared re-estimation error averaged over all the samples. This option is more time-consuming.

method2

Selection criterion for selecting the ACP pivots (in the iterative selection algorithm):

  • 1 Local criterion: The next pivot is the sample which receives the largest estimation error (comparing the estimated and the true values).

  • 2 Global criterion: This option has not been coded as it would lead to huge computing times.

ranks1

Array of ranks of samples already selected as exact pivots. It may be empty.

ranks2

Array of ranks of samples already selected as ACP pivots. It may be empty.

verbose

Verbose option

Value

A list containing:

The ranks are counted starting from 1.


[Package RGeostats version 14.0.10 Index]