simPopInternal: Internal functions for population simulation

In martinbryan/SUMMER: Small-Area-Estimation Unit/Area Models and Methods for Estimation in R

Internal functions for population simulation

Description

Functions for calculating valuable quantities and for drawing from important distributions for population simulation.

Usage

getExpectedNperEA(easpa, popMat, level = c("grid", "EA"), pixelIndexMat = NULL)

getSortIndices(
  i,
  urban = TRUE,
  popMat,
  stratifyByUrban = TRUE,
  validationPixelI = NULL
)

rStratifiedMultnomial(n, popMat, easpa, stratifyByUrban = TRUE)

rStratifiedMultnomialBySubarea(
  n,
  popMat,
  easpa,
  stratifyByUrban = TRUE,
  poppsub = NULL,
  min1PerSubarea = TRUE,
  minSample = 1
)

rMyMultinomial(
  n,
  i,
  stratifyByUrban = TRUE,
  urban = TRUE,
  popMat = NULL,
  easpa = NULL,
  min1PerSubarea = FALSE,
  method = c("mult1", "mult", "indepMH"),
  minSample = 1
)

rMyMultinomialSubarea(
  n,
  i,
  easpsub,
  stratifyByUrban = TRUE,
  urban = TRUE,
  popMat = NULL
)

rmultinom1(
  n = 1,
  size,
  prob,
  maxSize = 8000 * 8000,
  method = c("mult1", "mult", "indepMH"),
  verbose = FALSE,
  minSample = 100,
  maxExpectedSizeBeforeSwitch = 1000 * 1e+07,
  init = NULL,
  burnIn = floor(n/4),
  filterEvery = 10,
  zeroProbZeroSamples = TRUE,
  allowSizeLessThanK = FALSE
)

sampleNMultilevelMultinomial(
  nDraws = ncol(pixelIndexMat),
  pixelIndexMat = NULL,
  urbanMat = NULL,
  areaMat = NULL,
  easpaList,
  popMat,
  stratifyByUrban = TRUE,
  verbose = TRUE,
  returnEAinfo = FALSE,
  minHHPerEA = 25,
  fixHHPerEA = NULL,
  fixPopPerHH = NULL
)

sampleNMultilevelMultinomialFixed(
  clustersPerPixel,
  nDraws = ncol(pixelIndices),
  pixelIndices = NULL,
  urbanVals = NULL,
  areaVals = NULL,
  easpa,
  popMat,
  stratifyByUrban = TRUE,
  verbose = TRUE
)

Arguments

easpa	Census frame. See simPopCustom for details
popMat	data.frame of pixellated grid of population densities. See simPopCustom for details
level	Whether to calculate results at the integration grid or EA level
pixelIndexMat	Matrix of pixel indices associated with each EA and draw. Not required by getExpectedNperEA unless level == "EA"
i	Index
urban	If TRUE, calculate only for urban part of the area. If FALSE, for only rural part
stratifyByUrban	whether or not to stratify calculations by urban/rural classification
validationPixelI	CURRENTLY FOR TESTING PURPOSES ONLY a set of indices of pixels for which we want to simulate populations (used for pixel level validation)
n	Number of samples
poppsub	Population per subarea. See simPopCustom for details
min1PerSubarea	Whether or not to ensure there is at least 1 EA per subarea. See simPopCustom for details
minSample	The minimum number of samples per 'chunk' of samples for truncated multinomial sampling. Defaults to 1
method	If min1PerSubarea is TRUE, the sampling method for the truncated multinomial to use with rmulitnom1. rmultinom1 automatically switches between them depending on the number of expected samples. The methods are: mult1 rejection sampling from multinomial plus 1 in each category mult rejection sampling from multinomial if any category has zero count indepMH independent Metropolis-Hastings using multinomial plus 1 distribution as proposal
easpsub	This could either be total EAs per subarea, or subarea crossed with urban or rural if stratifyByUrban is TRUE
size	Multinomial size parameter. See rmultinom
prob	Multinomial probability vector parameter. See rmultinom
maxSize	The maximum number of elements in a matrix drawn from the proposal distribution per sample chunk.
verbose	Whether to print progress as the function proceeds
maxExpectedSizeBeforeSwitch	Max expected number of samples / k, the number of categories, before switching method
init	Initial sample if method is 'indepMH'
burnIn	Number of initial samples before samples are collected if method is 'indepMH'
filterEvery	Store only every filterEvery samples if method is i'indepMH'
zeroProbZeroSamples	If TRUE, set samples for parts of prob vector that are zero to zero. Otherwise they are set to one.
allowSizeLessThanK	If TRUE, then if size < the number of categories (k), returns matrix where each column is vector of size ones and k - size zeros. If FALSE, throws an error if size < k
nDraws	Number of draws
urbanMat	Matrix of urbanicities associated with each EA and draw
areaMat	Matrix of areas associated with each EA and draw
easpaList	A list of length n with each element being of the format of easpa giving the number of households and EAs per stratum. It is assumed that the number of EAs per stratum is the same in each list element. If easpaList is a data frame, number of households per stratum is assumed constant
returnEAinfo	Whether a data frame at the EA level is desired
minHHPerEA	The minimum number of households per EA (defaults to 25, since that is the number of households sampled per DHS cluster)
fixHHPerEA	If not NULL, the fixed number of households per EA
fixPopPerHH	If not NULL, the fixed target population per household
clustersPerPixel	CURRENTLY FOR TESTING PURPOSES ONLY a vector of length nIntegrationPoints specifying the number of clusters per pixel if they are fixed
pixelIndices	A nEA x n matrix of pixel indices associated with each EA per simulation/draw
urbanVals	A nEA x n matrix of urbanicities associated with each EA per simulation/draw
areaVals	A nEA x n matrix of area names associated with each EA per simulation/draw

Functions

• getExpectedNperEA(): Calculates expected denominator per enumeration area.

• getSortIndices(): For recombining separate multinomials into the draws over all grid points

• rStratifiedMultnomial(): Gives nIntegrationPoints x n matrix of draws from the stratified multinomial with values corresponding to the value of |C^g| for each pixel, g (the number of EAs/pixel)

• rStratifiedMultnomialBySubarea(): Gives nIntegrationPoints x n matrix of draws from the stratified multinomial with values

• rMyMultinomial():

• rMyMultinomialSubarea():

• rmultinom1(): Random (truncated) multinomial draws conditional on the number of each type being at least one

• sampleNMultilevelMultinomial(): Take multilevel multinomial draws first from joint distribution of number of households per EA given the total per stratum, and then from the joint distribution of the total target population per household given the total per stratum

• sampleNMultilevelMultinomialFixed(): Same as sampleNMultilevelMultinomial, except the number of EAs per pixel is fixed

martinbryan/SUMMER documentation built on April 10, 2024, 5:03 a.m.