fpbb_synth_pops: Create synthetic populations with FPBB.

In awmercer/bestimate: Bayesian Finite Population Survey Estimation With FPBB and BART

Description

Create synthetic populations with FPBB.

Usage

fpbb_synth_pops(weights, L = 2, N = length(weights) * 2,
  return_weights = TRUE)

Arguments

weights	Vector of weights for each unit in the reference sample. These determine the initial probabilities of selection for the weighted Polya urn scheme used to generate each synthetic population.
L	Number of synthetic populations to create.
N	Size of the synthetic populations. In theory this is the size of the full population but in practice only needs to be 10-20 times as large as the reference sample.
return_weights	Logical flag indicating whether the synthetic populations should be returned as weights or indices. If TRUE, output will contain the number of times each unit in the reference sample appears in the synthetic population. If FALSE output will be a vector containing the indices of each unit in the order they were selected. This permits the use of synthetic populations with procedure that cannot handle weights.

Value

Returns a data frame with one column for each synthetic population. If return_weights = TRUE each column will be a vector containing the number of times a unit appears in the synthetic population. If return_weights = FALSE each column will be a vector containing the row number of each unit in the reference sample in the order it was selected.

awmercer/bestimate documentation built on May 22, 2019, 8:50 p.m.