sdcMicro (version 4.1.0)

pram: Post RAndomisation Method (PRAM)

Description

PRAM is a probabilistic, perturbative method which can be applied on categorical variables.

Usage

pram(obj, pramVar=NULL,pd=0.8, alpha=0.5)

Arguments

obj

a numeric vector or factor, data.frame or object of class sdcMicroObj

pramVar

name of the pramVar that should be prammed

pd

minimum diagonal entries for the generated transition matrix P. Either a vector of length 1 or a vector of length ( number of categories ).

alpha

amount of perturbation for the invariant Pram method

Value

• a modified sdcMicroObj object or the following list
• x
  original vector
• xpramed
  the perturbed vector
• pd
  randomly generated diagonal entry of the P (between original pd and 1)
• Rs
  invariant transition matrix
• alpha
  amount of perturbation for the invariant Pram method

Details

The method is implemented exactly as described in the citation in the references. First a transition matrix is created in that way, that the diagonal entries of a matrix P are random numbers between pd and 1. The remaining entries of the matrix are generated such that the rowSums of the matrix is 1. Then a invariant transition matrix is generated.

References

Shlomo, Natalie and de Waal, Ton (2006) Protection of Micro-data Subject to Edit Constraints Against Statistical Disclosure. Southampton, UK, Southampton Statistical Sciences Research Institute, 36pp. (S3RI Methodology Working Papers, M06/16)

Examples

set.seed(123)
x <- factor(sample(1:4, 250, replace=TRUE))
pr1 <- pram(x)
length(which(pr1$x == x))
x2 <- as.factor(sample(1:4, 250, replace=TRUE))
length(which(pram(x2)$x == x2))

data(free1)
free1 <- data.frame(free1)
free1[,"MARSTAT"] <- as.factor(free1[,"MARSTAT"])
marstatPramed <- pram(free1[,"MARSTAT"])  

## for objects of class sdcMicro:
data(testdata2)
sdc <- createSdcObj(testdata2,
  keyVars=c('urbrur','roof','walls','sex'), 
  pramVars=c('water','electcon','relat'),
  numVars=c('expend','income','savings'), weightVar='sampling_weight')
sdc <- pram(sdc, pramVar="water")