Sparameters.default.f: Default synthetisation parameters based on variable names

In DanielBonnery/BigSyn: X

Description

Default synthetisation parameters based on variable names

Usage

Sparameters.default.f(
  ref.table,
  asis = NULL,
  notpredictor = NULL,
  variables = Sparameters.variables.reorder.default(names(ref.table)),
  predictors.matrix = predictor.matrix.default(variables)[!is.element(variables, asis),
    !is.element(variables, notpredictor)],
  splittingvar = NULL,
  moresplits = NULL,
  preferredmethod = "ctree",
  splithreshold = 100,
  defaultsynparameters = list(smoothing = "none", importance = TRUE, keep.forest = TRUE,
    minbucket = 30)
)

Arguments

ref.table	a dataframe
asis	a vector of character strings, indicating which variables to keep as is.
notpredictor	a vector of character strings, indicating which variables are not supposed to be used as predictors.
variables	a vector of character strings, indicating the variables to synthesize. Order is important.
predictors.matrix	a predictor matrix. Number of rows is the number of variables to synthesize, number of columns is all the variables from ref.table
moresplits	an object of class moresplist (not defined yet)
preferredmethod:	"rf" for random forest or "ctree" for classification tree
defaultparameters	a list indicating default parameters for synthpop synthesisation functions, for example ntree=5, smoothing="none"

Details

creates default synthetisation parameters Some rules: parents variable are potential predictors of their children, synthetisation is conditional to missingindicators, synthetisation is conditional to presence in cell

Examples

data(TtableA)
ATtableA<-augmentT_f(TtableA,variablespct="AA.cont1",variablesmax="AA.present")
ref.table<-ATtableA
Spa<-Sparameters.default.f(ref.table=ATtableA)
names(Spa)<-lapply(Spa,function(x){x$variable})
Spa$AA.present_La_Lb
Spa$AA.cont1_La_Lb

DanielBonnery/BigSyn documentation built on June 28, 2020, 7:18 p.m.