MaskJointDensity
Masking, Unmasking and Restoring Confidential Data

Package index
Search the MaskJointDensity package
Vignettes
Functions
33
Source code
28
Man pages
32
Home
/
CRAN
/
MaskJointDensity
/
MaskJointDensity-package: Masking, unmasking and restoring confidential data.

MaskJointDensity-package: Masking, unmasking and restoring confidential data.
In MaskJointDensity: Masking, Unmasking and Restoring Confidential Data

Description Details Author(s) References Examples

Description

This package has 3 key functionalities. It is able to mask confidential data using multiplicative noise. It is able to unmask this data while still preserving confidentiality. It is able to calculate the numerical joint density function of the original data from the unmasked data, as well as obtaining a sample from the marginal density functions of the unmasked data. The final results are a reasonable approximation to the original data for the purposes of analysis.

Details

Package: MaskJointDensity
Type: Package
Version: 1
Date: 2018-05-05
License: GPL-2

The Data Provider obtains confidential data, then masks it and writes it. The End-User unmasks this data then can obtain the joint density function of the unmasked variables and samples from the marginal density of them in order to mimic having the confidential data for the purpose of statistical analysis.

Author(s)

Yan-Xia Lin Luke Mazur Jordan Morris

Maintainer: Luke Mazur <lm810@uowmail.edu.au>

References

Yan-Xia Lin, Luke Mazur, Rathin Sarathy, Krishnamurty Muralidhar Statistical Information Recovery from Multivariate Noise-Multiplied Data, a Computational Approach Transactions on Data Privacy 11:1 (2018) 23 - 45 http://www.tdp.cat/issues16/abs.a271a17.php

Examples

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
# Data Provider
set.seed(100)
data(FPI2002Data)
OriginalVar01 <- log(FPI2002Data$FPIflow2)
OriginalVar02 <- log(FPI2002Data$gdp_o)  
OriginalVar03 <- log(FPI2002Data$gdp_d) 
OriginalVar04 <- log(FPI2002Data$distw) 

testN <- nrow(FPI2002Data)

noise<-rmulti(n=testN, mean=c(80, 100), sd=c(5,3), p=c(0.6, 0.4))
xmask<-mask(vectorToBeMasked=OriginalVar01, noisefile=file.path(tempdir(),"noise1.bin"),
noise=noise, lowerBoundAsGivenByProvider=min(OriginalVar01),
upperBoundAsGivenByProvider=max(OriginalVar01))
write(xmask$ystar, file.path(tempdir(),"xstar1.dat"))

noise<-rmulti(n=testN, mean=c(60, 90), sd=c(5,3), p=c(0.4, 0.6))
xmask<-mask(vectorToBeMasked=OriginalVar02, noisefile=file.path(tempdir(),"noise2.bin"),
noise=noise, lowerBoundAsGivenByProvider=min(OriginalVar02),
upperBoundAsGivenByProvider=max(OriginalVar02))
write(xmask$ystar, file.path(tempdir(),"xstar2.dat"))

noise<-rmulti(n=testN, mean=c(40, 110), sd=c(5,3), p=c(0.6, 0.4))
xmask<-mask(vectorToBeMasked=OriginalVar03, noisefile=file.path(tempdir(),"noise3.bin"),
noise=noise, lowerBoundAsGivenByProvider=min(OriginalVar03),
upperBoundAsGivenByProvider=max(OriginalVar03))
write(xmask$ystar, file.path(tempdir(),"xstar3.dat"))

noise<-rmulti(n=testN, mean=c(70, 100), sd=c(5,3), p=c(0.4, 0.6))
xmask<-mask(vectorToBeMasked=OriginalVar04, noisefile=file.path(tempdir(),"noise4.bin"),
noise=noise, lowerBoundAsGivenByProvider=min(OriginalVar04),
upperBoundAsGivenByProvider=max(OriginalVar04))
write(xmask$ystar, file.path(tempdir(),"xstar4.dat"))

# End-User

xstar1 <- scan(file.path(tempdir(),"xstar1.dat"))
Unmasked01 <- unmask(maskedVectorToBeUnmasked=xstar1,
noisefile=file.path(tempdir(),"noise1.bin"))

xstar2 <- scan(file.path(tempdir(),"xstar2.dat"))
Unmasked02 <- unmask(maskedVectorToBeUnmasked=xstar2,
noisefile=file.path(tempdir(),"noise2.bin"))

xstar3 <- scan(file.path(tempdir(),"xstar3.dat"))
Unmasked03 <- unmask(maskedVectorToBeUnmasked=xstar3,
noisefile=file.path(tempdir(),"noise3.bin"))

xstar4 <- scan(file.path(tempdir(),"xstar4.dat"))
Unmasked04 <- unmask(maskedVectorToBeUnmasked=xstar4,
noisefile=file.path(tempdir(),"noise4.bin"))

# If the Data Provider has given the End-User the mean, standard deviation,
# or correlation matrix of the original variables

listOfMeansOfOriginalVariables<-list(mean(OriginalVar01), mean(OriginalVar02),
mean(OriginalVar03), mean(OriginalVar04))
listOfStandardDeviationsOfOriginalVariables<-list(sd(OriginalVar01), sd(OriginalVar02), 
sd(OriginalVar03), sd(OriginalVar04))
matrixOfCorrelationsOfOriginalVariables<-cor(cbind(OriginalVar01, OriginalVar02,
OriginalVar03, OriginalVar04))

test1 <- getSampleBasedOnUnmaskedData(maskedVectors =
list(xstar1, xstar2, xstar3, xstar4),
unmaskedVectors = list(Unmasked01$unmaskedVariable, 
Unmasked02$unmaskedVariable, Unmasked03$unmaskedVariable, 
Unmasked04$unmaskedVariable),
mu = listOfMeansOfOriginalVariables,
s = listOfStandardDeviationsOfOriginalVariables,  
rho_X = matrixOfCorrelationsOfOriginalVariables,
verbose = 2,
size = 1000)

# If the Data Provider has not, then these are estimated

test2 <- getSampleBasedOnUnmaskedData(meansOfNoises = 
list(Unmasked01$meanOfNoise, Unmasked02$meanOfNoise, 
Unmasked03$meanOfNoise, Unmasked04$meanOfNoise),
meansOfSquaredNoises = list(Unmasked01$meanOfSquaredNoise,
Unmasked02$meanOfSquaredNoise, Unmasked03$meanOfSquaredNoise,
Unmasked04$meanOfSquaredNoise),
maskedVectors = list(xstar1, xstar2, xstar3, xstar4),
unmaskedVectors = list(Unmasked01$unmaskedVariable, 
Unmasked02$unmaskedVariable, Unmasked03$unmaskedVariable, 
Unmasked04$unmaskedVariable),
verbose = 2,
size = 1000)


## alternatively - using batch versions

# Data Provider
set.seed(100)
data(FPI2002Data)
OriginalVar01 <- log(FPI2002Data$FPIflow2)
OriginalVar02 <- log(FPI2002Data$gdp_o)  
OriginalVar03 <- log(FPI2002Data$gdp_d) 
OriginalVar04 <- log(FPI2002Data$distw) 

testN <- nrow(FPI2002Data)

noise1<-rmulti(n=testN, mean=c(80, 100), sd=c(5,3), p=c(0.6, 0.4))
noise2<-rmulti(n=testN, mean=c(60, 90), sd=c(5,3), p=c(0.4, 0.6))
noise3<-rmulti(n=testN, mean=c(40, 110), sd=c(5,3), p=c(0.6, 0.4))
noise4<-rmulti(n=testN, mean=c(70, 100), sd=c(5,3), p=c(0.4, 0.6))

maskBatchOut <- maskBatch(listOfVectorsToBeMasked=
list(OriginalVar01,OriginalVar02,OriginalVar03,OriginalVar04),
listOfNoisefiles=list(file.path(tempdir(),"noise1.bin"),file.path(tempdir(),"noise2.bin"),
file.path(tempdir(),"noise3.bin"),file.path(tempdir(),"noise4.bin")),
listOfNoises=list(noise1,noise2,noise3,noise4),
listOfLowerBoundsAsGivenByProvider=
list(min(OriginalVar01),min(OriginalVar02),min(OriginalVar03),min(OriginalVar04)),
listofUpperBoundsAsGivenByProvider=
list(max(OriginalVar01),max(OriginalVar02),max(OriginalVar03),max(OriginalVar04)),
maxorder = 100, EPS = 1e-06)

dataFileNames <- list(file.path(tempdir(),"xstar1.dat"),file.path(tempdir(),"xstar2.dat"),
file.path(tempdir(),"xstar3.dat"),file.path(tempdir(),"xstar4.dat"))

for(i in 1:length(maskBatchOut)) {
write((maskBatchOut[[i]])$ystar, dataFileNames[[i]])
}


# End-User

dataFileNames <- list(file.path(tempdir(),"xstar1.dat"),file.path(tempdir(),"xstar2.dat"),
file.path(tempdir(),"xstar3.dat"),file.path(tempdir(),"xstar4.dat")) 
numberOfFiles <- length(dataFileNames)

maskedVectors <- list()
for(i in 1:numberOfFiles) {
maskedVectors[[i]] <- scan(dataFileNames[[i]]) 
}

# If the Data Provider has given the End-User the mean, standard deviation,
# or correlation matrix of the original variables
listOfMeansOfOriginalVariables<-list(mean(OriginalVar01), mean(OriginalVar02),
mean(OriginalVar03), mean(OriginalVar04))
listOfStandardDeviationsOfOriginalVariables<-list(sd(OriginalVar01), sd(OriginalVar02), 
sd(OriginalVar03), sd(OriginalVar04))
matrixOfCorrelationsOfOriginalVariables<-cor(cbind(OriginalVar01, OriginalVar02,
OriginalVar03, OriginalVar04))

unmaskAndGetSampleOut1 <- unmaskAndGetSampleBatch(listOfMaskedVectorsToBeUnmasked=maskedVectors,
                                    listOfNoisefiles=
                                    list(file.path(tempdir(),"noise1.bin"),
                                    file.path(tempdir(),"noise2.bin"),
                                    file.path(tempdir(),"noise3.bin"),
                                    file.path(tempdir(),"noise4.bin")),
                                    mu=listOfMeansOfOriginalVariables, 
                                    s=listOfStandardDeviationsOfOriginalVariables, 
                                    rho_X=matrixOfCorrelationsOfOriginalVariables,
                                    cores = 1, size=1000,
                                    verbose = 2,
                                    onlyUnmasked = FALSE)


# If the Data Provider has not, then these are estimated

unmaskAndGetSampleOut2 <- unmaskAndGetSampleBatch(listOfMaskedVectorsToBeUnmasked=maskedVectors,
                                    listOfNoisefiles=
                                    list(file.path(tempdir(),"noise1.bin"),
                                    file.path(tempdir(),"noise2.bin"),
                                    file.path(tempdir(),"noise3.bin"),
                                    file.path(tempdir(),"noise4.bin")),
                                    cores = 1, size=1000,
                                    verbose = 2,
                                    onlyUnmasked = FALSE)

MaskJointDensity documentation built on May 2, 2019, 8:28 a.m.

Related to MaskJointDensity-package in MaskJointDensity...

MaskJointDensity index
Package overview

R Package Documentation

rdrr.io home R language documentation Run R code online

Browse R Packages

CRAN packages Bioconductor packages R-Forge packages GitHub packages

We want your feedback!

Note that we can't provide technical support on individual packages. You should contact the package authors for that.
GitHub issue tracker
ian@mutexlabs.com
Personal blog

 
Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.

Close