Simulation of Synthetic Populations for Survey Data Considering Auxiliary Information

Tools and methods to simulate populations for surveys based on auxiliary data. The tools include model-based methods, calibration and combinatorial optimization algorithms. The package was developed with support of the International Household Survey Network, DFID Trust Fund TF011722 and funds from the World bank.


News

Changes in simPop version 0.2.15

  • Improvements in samp() and pop(). If argument 'var' is NULL (default), the complete dataset is returned.

Changes in simPop version 0.2.14

  • Warn the user about missing values in variables used in the model in simContinuous()
  • additional argument imputeMissings (default=FALSE) in simContinuous()
  • do not impute missing variables in population data

Changes in simPop version 0.2.13

  • bug in simContinuous solved
  • simStructure now works also when no stratification variable is specified

Changes in simPop version 0.2.12

  • adding some utility functions for the sga-project that are not exported
  • improve simContinuous() in order to allow to specify the model that should be used for estimation.
  • improve simCategorical() in order to allow to specify the model that should be used for estimation.
  • problem fixed for missing values in variables not used for modelling.

Changes in simPop version 0.2.11

  • new methods samp, samp<-,pop,and pop<- to extract/modify variables in sample/population on a simPopObj
  • adjust manageSimPopObj() to use these methods

Changes in simPop version 0.2.10

  • improvements in simContinous(): new parameters 'byHousehold' and 'basicOnly'

Changes in simPop version 0.2.9

  • manageSynthPopObj to manageSimPopObj
  • bug in calibPop.cpp solved (segmentation fault)

Changes in simPop version 0.2.8

  • random numbers in C++ instead of using R

Changes in simPop version 0.2.6

  • fix clang-compiler warning

Changes in simPop version 0.2.5

  • multiple fixes and performance improvements

Changes in simPop version 0.2.4

  • feature: generate colored mosaic plots in spMosaic()
  • fix: missing argument and code-mismatch in simRelation()

Changes in simPop version 0.2.3

  • fix in simContinuous() reported by Oliver

Changes in simPop version 0.2.2

  • new parameter 'nr_cpus' for calibPop(), simCategorical(), simContinuous() and simRelation() manually defining how many cpus should be used
  • updated man-pages
  • new helper-function 'parallelParameters()' returing a list with parameters 'have_win', 'parallel' and 'nr_cores'
  • fixed a bug in simCategorical if only one additional variable is simulated

Changes in simPop version 0.2.1

  • fixed a bug in simContinuous (win only) reported by olivier.

Changes in simPop version 0.2.0

  • version drop for final delivery

Changes in simPop version 0.1.9

  • allow more userfriendly input for calibSample()
  • calibSample() can be applied to objects of class 'simPopObj'
  • updated example for ipu
  • whipple index (original and modified) included
  • sprague index included
  • simInitSpatial() to generate districts from a broader region
  • dataObj-methods for covWt() and corWt()
  • dataObj-methods for meanWt() and varWt()
  • method addWeights<- to modify sampling weights based on output from calibSample()
  • dataObj-method for quantileWt()
  • updated default parameters for calibPop(). Previous choices for starting temperature were much to high and too many worse solution were accepted.
  • fixed example for calibPop()
  • new accessor/set functions sampleData, sampleObj, sampleObj<-, popData, popObj, popObj<-, tableObj with updated man-page

Changes in simPop version 0.1.8

  • remove S3-class spBwplot() -> only default method was implemented. Removed option to supply a list of populations; rewritten to use 'simPopObj'-objects

  • fixed man files for spBwplot() and spBwplotStats()

  • updated calibPop() so that the function uses and returns objects of class 'simPopObj'

  • auxData()-methods to query/set 'sample' slot in 'simPopObj'-objects

Changes in simPop version 0.1.7

  • parallel-computing on windows-platforms using doParallel-package

  • removed spTable.formula method (did not work correctly, anyway)

  • parallel processing on windows using doParallel-package and foreach with %dopar%

  • fixed a problem in calibPop.cpp where export of params-object in a list was not done correctly

  • refactorization of calibPop.R to allow parallel computing

Changes in simPop version 0.1.6

  • removed method 'ctree' from simCategorical() -> perhaps add it later

  • removed argument 'basic' from simCategorical(). The function now uses variables listed in slot @basicHHvars of the input object of class 'simPopObj'.

  • made check that slot basicHHvars after simStructure() contains at least one variable.

  • new utility-function manageSynthPopObj() to get/set variables in objects of class 'simPopObj'

  • updated a lot of Rd-Files with (now) working examples

  • updated simEUSILC()

Changes in simPop version 0.1.5

  • combined classes 'sampleObj' and 'popObj' into 'dataObj'

  • renamed specify_sample() to specifyInput()

  • adjusted NAMESPACE files, documentation and functions to reflect these changes

  • Parallelization of simContinous() on non-windows platforms

  • porting simContinous() to use new class 'simPopObj' as input

Changes in simPop version 0.1.4

  • starting to use new classes 'sampleObj', 'popObj' and 'simPopObj' for the entire package

  • various fixes and improvements such as a new c++ implementation of calibVars()

  • cleanup dependencies in DESCRIPTION and NAMESPACE

  • remove method 'liblinear' from simCategorical() - has never really worked

Changes in simPop version 0.1.3

  • parallelize simRelevation() on non-windows platforms

  • c++ algorithm for simulated annealing in calibPop()

  • multiple commits to fix documentation and cran-notes/warnings when checking

Changes in simPop version 0.1.2

  • temporarily add explicit parallel-option to calibPop() for testing-purposes

  • more efficient implementation of auxiliary-function resample() used in calibPop()

  • changed default setting for auxiliary-variable 'factor' in calibPop() which now allows the algorithm to terminate in an accceptable time

  • added R/zzz.R to get current package version on loading the package

Changes in simPop version 0.1.1

  • more efficient c++ implementation of iterative proportional updating

  • multiple imports from package simPopulation

  • reorganizations (simAnnealing -> simPop,...)

Changes in simPop version 0.1.0

  • code largely based on version 0.4.1 of simPopulation

  • use package parallel if possible for some functions (non-windows platforms)

  • first version of simulated annealing

  • first (very rough) version of IPU (iterative proportional updating)

  • remove vignette from simPopulation

  • update Citation

  • new methods for simCategorical (naivebayes, ctree and liblinear)

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.

install.packages("simPop")

0.6.0 by Matthias Templ, 4 months ago


Browse source code at https://github.com/cran/simPop


Authors: Bernhard Meindl, Matthias Templ, Andreas Alfons, Alexander Kowarik, and with contributions from Mathieu Ribatet


Documentation:   PDF Manual  


Task views: Official Statistics & Survey Methodology


GPL (>= 2) license


Imports laeken, plyr, MASS, Rcpp, e1071, parallel, nnet, doParallel, foreach, colorspace, VIM, methods, party, EnvStats, fitdistrplus, ranger

Depends on lattice, vcd, data.table

Suggests haven, microbenchmark, stringr

Linking to Rcpp


See at CRAN