Package 'sdcMicro'

Description

specify variables that are linked to a key variable. This results in all suppressions of the key-variable being also applied on the corresponding 'ghost'-variables.

Usage

addGhostVars(obj, keyVar, ghostVars)

Arguments

Value

a modified sdcMicroObj-class object.

Author(s)

Bernhard Meindl

References

Templ, M. Statistical Disclosure Control for Microdata: Methods and Applications in R. Springer International Publishing, 287 pages, 2017. ISBN 978-3-319-50272-4. doi:10.1007/978-3-319-50272-4 doi:10.1007/978-3-319-50272-4

Examples

data(testdata2)
sdc <- createSdcObj(testdata2,
  keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
  numVars=c('expend','income','savings'), w='sampling_weight')
## we want to link the anonymization status of key variabe 'urbrur' to 'hhcivil'
sdc <- addGhostVars(sdc, keyVar="urbrur", ghostVars=c("hhcivil"))
## we want to link the anonymization status of key variabe 'roof' to 'represent'
sdc <- addGhostVars(sdc, keyVar="roof", ghostVars=c("represent"))

Description

Various methods for adding noise to perturb continuous scaled variables.

Usage

addNoise(obj, variables = NULL, noise = 150, method = "additive", ...)

Arguments

Details

If ‘obj’ is of class sdcMicroObj-class, all continuous key variables are selected per default. If ‘obj’ is of class “data.frame” or “matrix”, the continuous variables have to be specified.

Method ‘additive’ adds noise completely at random to each variable depending on its size and standard deviation. ‘correlated’ and method ‘correlated2’ adds noise and preserves the covariances as described in R. Brand (2001) or in the reference given below. Method ‘restr’ takes the sample size into account when adding noise. Method ‘ROMM’ is an implementation of the algorithm ROMM (Random Orthogonalized Matrix Masking) (Fienberg, 2004). Method ‘outdect’ adds noise only to outliers. The outliers are identified with univariate and robust multivariate procedures based on a robust mahalanobis distances calculated by the MCD estimator.

Value

If ‘obj’ was of class sdcMicroObj-class the corresponding slots are filled, like manipNumVars, risk and utility.

If ‘obj’ was of class “data.frame” or “matrix” an object of class “micro” with following entities is returned:

Author(s)

Matthias Templ and Bernhard Meindl

References

Domingo-Ferrer, J. and Sebe, F. and Castella, J., “On the security of noise addition for privacy in statistical databases”, Lecture Notes in Computer Science, vol. 3050, pp. 149-161, 2004. ISSN 0302-9743. Vol. Privacy in Statistical Databases, eds. J. Domingo-Ferrer and V. Torra, Berlin: Springer-Verlag.

Ting, D. Fienberg, S.E. and Trottini, M. “ROMM Methodology for Microdata Release” Joint UNECE/Eurostat work session on statistical data confidentiality, Geneva, Switzerland, 2005, https://www.unece.org/fileadmin/DAM/stats/documents/ece/ces/ge.46/2005/wp.11.e.pdf

Ting, D., Fienberg, S.E., Trottini, M. “Random orthogonal matrix masking methodology for microdata release”, International Journal of Information and Computer Security, vol. 2, pp. 86-105, 2008.

Templ, M. and Meindl, B., Robustification of Microdata Masking Methods and the Comparison with Existing Methods, Lecture Notes in Computer Science, Privacy in Statistical Databases, vol. 5262, pp. 177-189, 2008.

Templ, M. New Developments in Statistical Disclosure Control and Imputation: Robust Statistics Applied to Official Statistics, Suedwestdeutscher Verlag fuer Hochschulschriften, 2009, ISBN: 3838108280, 264 pages.

Templ, M. and Meindl, B. and Kowarik, A.: Statistical Disclosure Control for Micro-Data Using the R Package sdcMicro, Journal of Statistical Software, 67 (4), 1–36, 2015. doi:10.18637/jss.v067.i04

Templ, M. Statistical Disclosure Control for Microdata: Methods and Applications in R. Springer International Publishing, 287 pages, 2017. ISBN 978-3-319-50272-4. doi:10.1007/978-3-319-50272-4

See Also

sdcMicroObj-class, summary.micro

Examples

data(Tarragona)

a1 <- addNoise(Tarragona)
a1


data(testdata)

# donttest because Examples with CPU time > 2.5 times elapsed time
testdata[, c('expend','income','savings')] <-
addNoise(testdata[,c('expend','income','savings')])$xm

## for objects of class sdcMicroObj:
data(testdata2)
sdc <- createSdcObj(testdata2,
  keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
  numVars=c('expend','income','savings'), w='sampling_weight')
sdc <- addNoise(sdc)

Description

calls microaggregation code from mu-argus. In case only one variable should be microaggregated and useOptimal is TRUE, Hansen-Mukherjee polynomial exact method is applied. In any other case, the Mateo-Domingo method is used.

Usage

argus_microaggregation(df, k, useOptimal = FALSE)

Arguments

Value

a list with two elements

• original: the originally provided input data

• microaggregated: the microaggregated data.frame

See Also

mu-Argus manual at https://github.com/sdcTools/manuals/raw/master/mu-argus/MUmanual5.1.pdf

Examples

mat <- matrix(sample(1:100, 50, replace=TRUE), nrow=10, ncol=5)
df <- as.data.frame(mat)
res <- argus_microaggregation(df, k=5, useOptimal=FALSE)

Description

argus_rankswap

Usage

argus_rankswap(df, perc)

Arguments

Value

a list with two elements

• original: the originally provided input data

• swapped: the data.frame containing the swapped values

See Also

mu-Argus manual at https://github.com/sdcTools/manuals/raw/master/mu-argus/MUmanual5.1.pdf

Examples

mat <- matrix(sample(1:100, 50, replace=TRUE), nrow=10, ncol=5)
df <- as.data.frame(mat)
res <- argus_rankswap(df, perc=10)

Description

Recomputation of Risk should be done after manual changing the content of an object of class sdcMicroObj

Usage

calcRisks(obj, ...)

Arguments

Details

By applying this function, the dislosure risk is re-estimated and the corresponding slots of an object of class sdcMicroObj are updated. This function mostly used internally to automatically update the risk after an sdc method is applied.

Value

a sdcMicroObj object with updated risk values

See Also

sdcMicroObj

Examples

data(testdata2)
sdc <- createSdcObj(testdata2,
  keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
  numVars=c('expend','income','savings'), w='sampling_weight')
sdc <- calcRisks(sdc)

Description

Small Toy Example Data set which was used by Sanz-Mateo et.al.

Format

The format is: int [1:13, 1:7] 10 12 17 21 9 12 12 14 13 15 ... - attr(*, "dimnames")=List of 2 ..$ : chr [1:13] "1" "2" "3" "4" ... ..$ : chr [1:7] "1" "2" "3" "4" ...

Examples

data(casc1)
casc1

Description

This test data set was obtained on July 27, 2000 using the public use Data Extraction System of the U.S. Bureau of the Census.

Format

A data frame sampled from year 1995 with 1080 observations on the following 13 variables.

AFNLWGT

Final weight (2 implied decimal places)

AGI

Adjusted gross income

EMCONTRB

Employer contribution for hlth insurance

FEDTAX

Federal income tax liability

PTOTVAL

Total person income

STATETAX

State income tax liability

TAXINC

Taxable income amount

POTHVAL

Total other persons income

INTVAL

Amt of interest income

PEARNVAL

Total person earnings

FICA

Soc. sec. retirement payroll deduction

WSALVAL

Amount: Total Wage and salary

ERNVAL

Business or Farm net earnings

Source

Public use file from the CASC project. More information on this test data can be found in the paper listed below.

References

Brand, R. and Domingo-Ferrer, J. and Mateo-Sanz, J.M., Reference data sets to test and compare SDC methods for protection of numerical microdata. Unpublished. https://research.cbs.nl/casc/CASCrefmicrodata.pdf

Examples

data(CASCrefmicrodata)
str(CASCrefmicrodata)

Description

createDat() returns dummy data to illustrate targeted record swapping. The generated data contain household ids ('hid'), geographic variables ('nuts1', 'nuts2', 'nuts3', 'lau2') as well as some other household or personal variables.

Usage

createDat(N = 10000)

Arguments

Value

'data.table' containing dummy data

See Also

recordSwap

Description

This is useful if the record IDs consist, for example, of a geo identifier and the household line number. This method can be used to create new, random IDs that cannot be reconstructed.

Usage

createNewID(obj, newID, withinVar)

Arguments

Value

an sdcMicroObj-class-object with updated slot origData

Description

Fast generation of (primitive) synthetic multivariate normal data.

Usage

dataGen(obj, ...)

Arguments

Details

Uses the cholesky decomposition to generate synthetic data with approx. the same means and covariances. For details see at the reference.

Value

the generated synthetic data.

Note

With this method only multivariate normal distributed data with approxiomately the same covariance as the original data can be generated without reflecting the distribution of real complex data, which are, in general, not follows a multivariate normal distribution.

Author(s)

Matthias Templ

References

Mateo-Sanz, Martinez-Balleste, Domingo-Ferrer. Fast Generation of Accurate Synthetic Microdata. International Workshop on Privacy in Statistical Databases PSD 2004: Privacy in Statistical Databases, pp 298-306.

See Also

sdcMicroObj-class, shuffle

Examples

data(mtcars)

cov(mtcars[,4:6])
cov(dataGen(mtcars[,4:6]))
pairs(mtcars[,4:6])
pairs(dataGen(mtcars[,4:6]))

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

Description

Distribute number of swaps across lowest hierarchy level according to a predefined swaprate. The swaprate is applied such that a single swap counts as swapping 2 households. Number of swaps are randomly rounded up or down, if needed, such that the total number of swaps is in coherence with the swaprate.
NOTE: This is an internal function used for testing the C++-function distributeDraws which is used inside the C++-function recordSwap().

Usage

distributeDraws_cpp(data, hierarchy, hid, swaprate, seed = 123456L)

Arguments

Description

Distribute 'totalDraws' using ratio/probability vector 'inputRatio' and randomly round each entry up or down such that the distribution results in an integer vector. Returns an integer vector containing the number of units in 'totalDraws' distributetd according to proportions in 'inputRatio'.
NOTE: This is an internal function used for testing the C++-function distributeRandom which is used inside the C++-function recordSwap().

Usage

distributeRandom_cpp(inputRatio, totalDraws, seed)

Arguments

Description

Distance-based disclosure risk estimation via standard deviation-based intervals around observations.

Usage

dRisk(obj, ...)

Arguments

Details

An interval (based on the standard deviation) is built around each value of the perturbed value. Then we look if the original values lay in these intervals or not. With parameter k one can enlarge or down scale the interval.

Value

The disclosure risk or/and the modified sdcMicroObj-class

Author(s)

Matthias Templ

References

see method SDID in Mateo-Sanz, Sebe, Domingo-Ferrer. Outlier Protection in Continuous Microdata Masking. International Workshop on Privacy in Statistical Databases. PSD 2004: Privacy in Statistical Databases pp 201-215.

Templ, M. Statistical Disclosure Control for Microdata: Methods and Applications in R. Springer International Publishing, 287 pages, 2017. ISBN 978-3-319-50272-4. doi:10.1007/978-3-319-50272-4

See Also

dUtility

Examples

data(free1)
free1 <- as.data.frame(free1)

m1 <- microaggregation(free1[, 31:34], method="onedims", aggr=3)
m2 <- microaggregation(free1[, 31:34], method="pca", aggr=3)
dRisk(obj=free1[, 31:34], xm=m1$mx)
dRisk(obj=free1[, 31:34], xm=m2$mx)
dUtility(obj=free1[, 31:34], xm=m1$mx)
dUtility(obj=free1[, 31:34], xm=m2$mx)

## for objects of class sdcMicro:
data(testdata2)
sdc <- createSdcObj(testdata2,
  keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
  numVars=c('expend','income','savings'), w='sampling_weight')
## this is already made internally: sdc <- dRisk(sdc)
## and already stored in sdc

Description

Distance-based disclosure risk estimation via robust Mahalanobis Distances.

Usage

dRiskRMD(obj, ...)

Arguments

Details

This method is an extension of method SDID because it accounts for the “outlyingness” of each observations. This is a quite natural approach since outliers do have a higher risk of re-identification and therefore these outliers should have larger disclosure risk intervals as observations in the center of the data cloud.

The algorithm works as follows:

1. Robust Mahalanobis distances are estimated in order to get a robust multivariate distance for each observation.

2. Intervals are estimated for each observation around every data point of the original data points where the length of the interval is defined/weighted by the squared robust Mahalanobis distance and the parameter $k$. The higher the RMD of an observation the larger the interval.

3. Check if the corresponding masked values fall into the intervals around the original values or not. If the value of the corresponding observation is within such an interval the whole observation is considered unsafe. So, we get a whole vector indicating which observation is save or not, and we are finished already when using method RMDID1).

4. For method RMDID1w: we return the weighted (via RMD) vector of disclosure risk.

5. For method RMDID2: whenever an observation is considered unsafe it is checked if $m$ other observations from the masked data are very close (defined by a parameter $k2$ for the length of the intervals as for SDID or RSDID) to such an unsafe observation from the masked data, using Euclidean distances. If more than $m$ points are in such a small interval, we conclude that this observation is “save”.

Value

The disclosure risk or the modified sdcMicroObj-class

Author(s)

Matthias Templ

References

Templ, M. and Meindl, B., Robust Statistics Meets SDC: New Disclosure Risk Measures for Continuous Microdata Masking, Lecture Notes in Computer Science, Privacy in Statistical Databases, vol. 5262, pp. 113-126, 2008.

Templ, M. New Developments in Statistical Disclosure Control and Imputation: Robust Statistics Applied to Official Statistics, Suedwestdeutscher Verlag fuer Hochschulschriften, 2009, ISBN: 3838108280, 264 pages.

See Also

dRisk

Examples

data(Tarragona)
x <- Tarragona[, 5:7]
y <- addNoise(x)$xm
dRiskRMD(x, xm=y)
dRisk(x, xm=y)

data(testdata2)
sdc <- createSdcObj(testdata2,
  keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
  numVars=c('expend','income','savings'), w='sampling_weight')

sdc <- dRiskRMD(sdc)

Description

dUtility() allows to compute different measures of data-utility based on various distances using original and perturbed variables.

Usage

dUtility(obj, ...)

Arguments

Details

The standardised distances of the perturbed data values to the original ones are measured. The following measures are available:

• ⁠"IL1⁠: sum of absolute distances between original and perturbed variables scaled by absolute values of the original variables

• ⁠"IL1s⁠: measures the absolute distances between original and perturbed ones, scaled by the standard deviation of original variables times the square root of 2.

• ⁠"eigen⁠; compares the eigenvalues of original and perturbed data

• ⁠"robeigen⁠; compares robust eigenvalues of original and perturbed data

Value

data utility or modified entry for data utility the sdcMicroObj.

Author(s)

Matthias Templ

References

for IL1 and IL1s: see Mateo-Sanz, Sebe, Domingo-Ferrer. Outlier Protection in Continuous Microdata Masking. International Workshop on Privacy in Statistical Databases. PSD 2004: Privacy in Statistical Databases pp 201-215.

Templ, M. and Meindl, B., ⁠Robust Statistics Meets SDC: New Disclosure Risk Measures for Continuous Microdata Masking⁠, Lecture Notes in Computer Science, Privacy in Statistical Databases, vol. 5262, pp. 113-126, 2008.

See Also

dRisk(), dRiskRMD()

Examples

data(free1)
free1 <- as.data.frame(free1)

m1 <- microaggregation(free1[, 31:34], method="onedims", aggr=3)
m2 <- microaggregation(free1[, 31:34], method="pca", aggr=3)
dRisk(obj=free1[, 31:34], xm=m1$mx)
dRisk(obj=free1[, 31:34], xm=m2$mx)
dUtility(obj=free1[, 31:34], xm=m1$mx)
dUtility(obj=free1[, 31:34], xm=m2$mx)
data(Tarragona)
x <- Tarragona[, 5:7]
y <- addNoise(x)$xm
dRiskRMD(x, xm=y)
dRisk(x, xm=y)
dUtility(x, xm = y, method = "IL1")
dUtility(x, xm = y, method = "IL1s")
dUtility(x, xm = y, method = "eigen")
dUtility(x, xm = y, method = "robeigen")

## for objects of class sdcMicro:
data(testdata2)
sdc <- createSdcObj(testdata2,
  keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
  numVars=c('expend','income','savings'), w='sampling_weight')
## this is already made internally, so you don't need to run this:
sdc <- dUtility(sdc)

Description

Data set obtained from the U.S. Energy Information Authority.

Format

A data frame with 4092 observations on the following 15 variables.

UTILITYID

UNIQUE UTILITY IDENTIFICATION NUMBER

UTILNAME

UTILITY NAME. A factor with levels 4-County Electric Power Assn Alabama Power Co Alaska Electric Appalachian Electric Coop Appalachian Power Co Arizona Public Service Co Arkansas Power & Light Co Arkansas Valley Elec Coop Corp Atlantic City Electric Company Baker Electric Coop Inc Baltimore Gas & Electric Co Bangor Hydro-Electric Co Berkeley Electric Coop Inc Black Hills Corp Blackstone Valley Electric Co Bonneville Power Admin Boston Edison Co Bountiful City Light & Power Bristol City of Brookings City of Brunswick Electric Member Corp Burlington City of Carolina Power & Light Co Carroll Electric Coop Corp Cass County Electric Coop Inc Central Illinois Light Company Central Illinois Pub Serv Co Central Louisiana Elec Co Inc Central Maine Power Co Central Power & Light Co Central Vermont Pub Serv Corp Chattanooga City of Cheyenne Light Fuel & Power Co Chugach Electric Assn Inc Cincinnati Gas & Electric Co Citizens Utilities Company City of Boulder City City of Clinton City of Dover City of Eugene City of Gillette City of Groton Dept of Utils City of Idaho Falls City of Independence City of Newark City of Reading City of Tupelo Water & Light D Clarksville City of Cleveland City of Cleveland Electric Illum Co Coast Electric Power Assn Cobb Electric Membership Corp Colorado River Commission Colorado Springs City of Columbus Southern Power Co Commonwealth Edison Co Commonwealth Electric Co Connecticut Light & Power Co Consolidated Edison Co-NY Inc Consumers Power Co Cornhusker Public Power Dist Cuivre River Electric Coop Inc Cumberland Elec Member Corp Dakota Electric Assn Dawson County Public Pwr Dist Dayton Power & Light Company Decatur City of Delaware Electric Coop Inc Delmarva Power & Light Co Detroit Edison Co Duck River Elec Member Corp Duke Power Co Duquesne Light Company East Central Electric Assn Eastern Maine Electric Coop El Paso Electric Co Electric Energy Inc Empire District Electric Co Exeter & Hampton Electric Co Fairbanks City of Fayetteville Public Works Comm First Electric Coop Corp Florence City of Florida Power & Light Co Florida Power Corp Fort Collins Lgt & Pwr Utility Fremont City of Georgia Power Co Gibson County Elec Member Corp Golden Valley Elec Assn Inc Grand Island City of Granite State Electric Co Green Mountain Power Corp Green River Electric Corp Greeneville City of Gulf Power Company Gulf States Utilities Co Hasting Utilities Hawaii Electric Light Co Inc Hawaiian Electric Co Inc Henderson-Union Rural E C C Homer Electric Assn Inc Hot Springs Rural El Assn Inc Houston Lighting & Power Co Huntsville City of Idaho Power Co IES Utilities Inc Illinois Power Co Indiana Michigan Power Co Indianapolis Power & Light Co Intermountain Rural Elec Assn Interstate Power Co Jackson Electric Member Corp Jersey Central Power&Light Co Joe Wheeler Elec Member Corp Johnson City City of Jones-Onslow Elec Member Corp Kansas City City of Kansas City Power & Light Co Kentucky Power Co Kentucky Utilities Co Ketchikan Public Utilities Kingsport Power Co Knoxville City of Kodiak Electric Assn Inc Kootenai Electric Coop, Inc Lansing Board of Water & Light Lenoir City City of Lincoln City of Long Island Lighting Co Los Angeles City of Louisiana Power & Light Co Louisville Gas & Electric Co Loup River Public Power Dist Lower Valley Power & Light Inc Maine Public Service Company Massachusetts Electric Co Matanuska Electric Assn Inc Maui Electric Co Ltd McKenzie Electric Coop Inc Memphis City of MidAmerican Energy Company Middle Tennessee E M C Midwest Energy, Inc Minnesota Power & Light Co Mississippi Power & Light Co Mississippi Power Co Monongahela Power Co Montana-Dakota Utilities Co Montana Power Co Moon Lake Electric Assn Inc Narragansett Electric Co Nashville City of Nebraska Public Power District Nevada Power Co New Hampshire Elec Coop, Inc New Orleans Public Service Inc New York State Gas & Electric Newport Electric Corp Niagara Mohawk Power Corp Nodak Rural Electric Coop Inc Norris Public Power District Northeast Oklahoma Electric Co Northern Indiana Pub Serv Co Northern States Power Co Northwestern Public Service Co Ohio Edison Co Ohio Power Co Ohio Valley Electric Corp Oklahoma Electric Coop, Inc Oklahoma Gas & Electric Co Oliver-Mercer Elec Coop, Inc Omaha Public Power District Otter Tail Power Co Pacific Gas & Electric Co Pacificorp dba Pacific Pwr & L Palmetto Electric Coop, Inc Pennsylvania Power & Light Co Pennyrile Rural Electric Coop Philadelphia Electric Co Pierre Municipal Electric Portland General Electric Co Potomac Edison Co Potomac Electric Power Co Poudre Valley R E A, Inc Power Authority of State of NY Provo City Corporation Public Service Co of Colorado Public Service Co of IN Inc Public Service Co of NH Public Service Co of NM Public Service Co of Oklahoma Public Service Electric&Gas Co PUD No 1 of Clark County PUD No 1 of Snohomish County Puget Sound Power & Light Co Rappahannock Electric Coop Rochester Public Utilities Rockland Electric Company Rosebud Electric Coop Inc Rutherford Elec Member Corp Sacramento Municipal Util Dist Salmon River Electric Coop Inc Salt River Proj Ag I & P Dist San Antonio City of Savannah Electric & Power Co Seattle City of Sierra Pacific Power Co Singing River Elec Power Assn Sioux Valley Empire E A Inc South Carolina Electric&Gas Co South Carolina Pub Serv Auth South Kentucky Rural E C C Southern California Edison Co Southern Nebraska Rural P P D Southern Pine Elec Power Assn Southwest Tennessee E M C Southwestern Electric Power Co Southwestern Public Service Co Springfield City of St Joseph Light & Power Co State Level Adjustment Tacoma City of Tampa Electric Co Texas-New Mexico Power Co Texas Utilities Electric Co Tri-County Electric Assn Inc Tucson Electric Power Co Turner-Hutchinsin El Coop, Inc TVA U S Bureau of Indian Affairs Union Electric Co Union Light Heat & Power Co United Illuminating Co Upper Cumberland E M C UtiliCorp United Inc Verdigris Valley Electric Coop Verendrye Electric Coop Inc Virginia Electric & Power Co Volunteer Electric Coop Wallingford Town of Warren Rural Elec Coop Corp Washington Water Power Co Watertown Municipal Utils Dept Wells Rural Electric Co West Penn Power Co West Plains Electric Coop Inc West River Electric Assn, Inc Western Massachusetts Elec Co Western Resources Inc Wheeling Power Company Wisconsin Electric Power Co Wisconsin Power & Light Co Wisconsin Public Service Corp Wright-Hennepin Coop Elec Assn Yellowstone Vlly Elec Coop Inc

STATE

STATE FOR WHICH THE UTILITY IS REPORTING. A factor with levels AK AL AR AZ CA CO CT DC DE FL GA HI IA ID IL IN KS KY LA MA MD ME MI MN MO MS MT NC ND NE NH NJ NM NV NY OH OK OR PA RI SC SD TN TX UT VA VT WA WI WV WY

YEAR

REPORTING YEAR FOR THE DATA

MONTH

REPORTING MONTH FOR THE DATA

RESREVENUE

REVENUE FROM SALES TO RESIDENTIAL CONSUMERS

RESSALES

SALES TO RESIDENTIAL CONSUMERS

COMREVENUE

REVENUE FROM SALES TO COMMERCIAL CONSUMERS

COMSALES

SALES TO COMMERCIAL CONSUMERS

INDREVENUE

REVENUE FROM SALES TO INDUSTRIAL CONSUMERS

INDSALES

SALES TO INDUSTRIAL CONSUMERS

OTHREVENUE

REVENUE FROM SALES TO OTHER CONSUMERS

OTHRSALES

SALES TO OTHER CONSUMERS

TOTREVENUE

REVENUE FROM SALES TO ALL CONSUMERS

TOTSALES

SALES TO ALL CONSUMERS

Source

Public use file from the CASC project.

References

Brand, R. and Domingo-Ferrer, J. and Mateo-Sanz, J.M., Reference data sets to test and compare SDC methods for protection of numerical microdata. Unpublished. https://research.cbs.nl/casc/CASCrefmicrodata.pdf

Examples

data(EIA)
head(EIA)

Description

Extract the manipulated data from an object of class sdcMicroObj-class

Usage

extractManipData(
  obj,
  ignoreKeyVars = FALSE,
  ignorePramVars = FALSE,
  ignoreNumVars = FALSE,
  ignoreGhostVars = FALSE,
  ignoreStrataVar = FALSE,
  randomizeRecords = "no"
)

Arguments

Value

a data.frame containing the anonymized data set

Author(s)

Alexander Kowarik, Bernhard Meindl

Examples

## for objects of class sdcMicro:
data(testdata2)
sdc <- createSdcObj(testdata,
  keyVars=c('urbrur','roof'),
  numVars=c('expend','income','savings'), w='sampling_weight')
sdc <- removeDirectID(sdc, var="age")
dataM <- extractManipData(sdc)

Description

Small synthetic data from Capobianchi, Polettini, Lucarelli

Format

A data frame with 8 observations on the following 8 variables.

Num1

a numeric vector

Key1

Key variable 1. A numeric vector

Num2

a numeric vector

Key2

Key variable 2. A numeric vector

Key3

Key variable 3. A numeric vector

Key4

Key variable 4. A numeric vector

Num3

a numeric vector

w

The weight vector. A numeric vector

Details

This data set is very similar to that one which are used by the authors of the paper given below. We need this data set only for demonstration effect, i.e. that the package provides the same results as their software.

Source

https://research.cbs.nl/casc/deliv/12d1.pdf

Examples

data(francdat)
francdat

Description

The public use toy demo data set from the mu-Argus software for SDC.

Format

The format is: num [1:4000, 1:34] 36 36 36 36 36 36 36 36 36 36 ... - attr(*, "dimnames")=List of 2 ..$ : NULL ..$ : chr [1:34] "REGION" "SEX" "AGE" "MARSTAT" ...

Details

Please, see at the link given below. Please note, that the correlation structure of the data is not very realistic, especially concerning the continuous scaled variables which drawn independently from are a multivariate uniform distribution.

Source

Public use file from the CASC project.

Examples

data(free1)
head(free1)

Description

Extract sample frequency counts (fk) or estimated population frequency counts (Fk)

Usage

freq(obj, type = "fk")

Arguments

Value

a vector containing sample frequencies or weighted frequencies

Author(s)

Bernhard Meindl

Examples

data(testdata)
sdc <- createSdcObj(testdata,
  keyVars=c('urbrur','roof','walls','relat','sex'),
  pramVars=c('water','electcon'),
  numVars=c('expend','income','savings'), w='sampling_weight')
head(freq(sdc, type="fk"))
head(freq(sdc, type="Fk"))

Description

Computation and estimation of the sample and population frequency counts.

Usage

freqCalc(x, keyVars, w = NULL, alpha = 1)

Arguments

Details

The function considers the case of missing values in the data. A missing value stands for any of the possible categories of the variable considered. It is possible to apply this function to large data sets with many (catergorical) key variables, since the computation is done in C.

freqCalc() does not support sdcMicro S4 class objects.

Value

Object from class freqCalc.

Author(s)

Bernhard Meindl

References

look e.g. in https://research.cbs.nl/casc/deliv/12d1.pdf Templ, M. Statistical Disclosure Control for Microdata Using the R-Package sdcMicro, Transactions on Data Privacy, vol. 1, number 2, pp. 67-85, 2008. https://www.tdp.cat/issues/abs.a004a08.php

Templ, M. New Developments in Statistical Disclosure Control and Imputation: Robust Statistics Applied to Official Statistics, Suedwestdeutscher Verlag fuer Hochschulschriften, 2009, ISBN: 3838108280, 264 pages.

Templ, M. Statistical Disclosure Control for Microdata: Methods and Applications in R. Springer International Publishing, 287 pages, 2017. ISBN 978-3-319-50272-4. doi:10.1007/978-3-319-50272-4 doi:10.1007/978-3-319-50272-4

Templ, M. and Meindl, B.: Practical Applications in Statistical Disclosure Control Using R, Privacy and Anonymity in Information Management Systems New Techniques for New Practical Problems, Springer, 31-62, 2010, ISBN: 978-1-84996-237-7.

See Also

indivRisk, measure_risk

Examples

data(francdat)

f <- freqCalc(francdat, keyVars=c(2,4,5,6),w=8)
f
f$freqCalc
f$fk
f$Fk
## with missings:
x <- francdat
x[3,5] <- NA
x[4,2] <- x[4,4] <- NA
x[5,6]  <- NA
x[6,2]  <- NA
f2 <- freqCalc(x, keyVars=c(2,4,5,6),w=8)
cbind(f2$fk, f2$Fk)

## test parameter 'alpha'
f3a <- freqCalc(x, keyVars=c(2,4,5,6), w=8, alpha=1)
f3b <- freqCalc(x, keyVars=c(2,4,5,6), w=8, alpha=0.5)
f3c <- freqCalc(x, keyVars=c(2,4,5,6), w=8, alpha=0.1)
data.frame(fka=f3a$fk, fkb=f3b$fk, fkc=f3c$fk)
data.frame(Fka=f3a$Fk, Fkb=f3b$Fk, Fkc=f3c$Fk)

Description

For strata defined by multiple variables (e.g. sex,age,country) one combined variable is generated.

Usage

generateStrata(df, stratavars, name)

Arguments

Value

The original data set with one new column.

Author(s)

Alexander Kowarik

Examples

x <- testdata
x <- generateStrata(x,c("sex","urbrur"),"strataIDvar")
head(x)

Description

extract information from sdcMicroObj-class-objects depending on argument type

Usage

get.sdcMicroObj(object, type)

Arguments

Value

a slot of a sdcMicroObj-class-object depending on argument type

Examples

sdc <- createSdcObj(testdata2,
  keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
  numVars=c('expend','income','savings'), w='sampling_weight')
sl <- slotNames(sdc)
res <- sapply(sl, function(x) get.sdcMicroObj(sdc, type=x))
str(res)

Description

Global recoding of variables

Usage

globalRecode(obj, ...)

Arguments

Details

If a labels parameter is specified, its values are used to name the factor levels. If none is specified, the factor level labels are constructed.

Value

the modified sdcMicroObj-class or a factor, unless labels = FALSE which results in the mere integer level codes.

Note

globalRecode can not be applied to vectors stored as factors from sdcMicro >= 4.7.0!

Author(s)

Matthias Templ and Bernhard Meindl

References

Templ, M. and Kowarik, A. and Meindl, B. Statistical Disclosure Control for Micro-Data Using the R Package sdcMicro. Journal of Statistical Software, 67 (4), 1–36, 2015. doi:10.18637/jss.v067.i04

Templ, M. Statistical Disclosure Control for Microdata: Methods and Applications in R. Springer International Publishing, 287 pages, 2017. ISBN 978-3-319-50272-4. doi:10.1007/978-3-319-50272-4 doi:10.1007/978-3-319-50272-4

See Also

cut

Examples

data(free1)
free1 <- as.data.frame(free1)

## application to a vector
head(globalRecode(free1$AGE, breaks=c(1,9,19,29,39,49,59,69,100), labels=1:8))
table(globalRecode(free1$AGE, breaks=c(1,9,19,29,39,49,59,69,100), labels=1:8))

## application to a data.frame
# automatic labels
table(globalRecode(free1, column="AGE", breaks=c(1,9,19,29,39,49,59,69,100))$AGE)

## calculation of brea-points using different algorithms
table(globalRecode(free1$AGE, breaks=6))
table(globalRecode(free1$AGE, breaks=6, method="logEqui"))
table(globalRecode(free1$AGE, breaks=6, method="equalAmount"))

## for objects of class sdcMicro:
data(testdata2)
sdc <- createSdcObj(testdata2,
  keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
  numVars=c('expend','income','savings'), w='sampling_weight')
sdc <- globalRecode(sdc, column="water", breaks=3)
table(get.sdcMicroObj(sdc, type="manipKeyVars")$water)

Join levels of a variables in an object of class sdcMicroObj-class or factor or data.frame

Description

If the input is an object of class sdcMicroObj-class, the specified factor-variable is recoded into a factor with less levels and risk-measures are automatically recomputed.

Usage

groupAndRename(obj, var, before, after, addNA = FALSE)

Arguments

Details

If the input is of class data.frame, the result is a data.frame with a modified column specified by var.

If the input is of class factor, the result is a factor with different levels.

Value

the modified sdcMicroObj-class

Author(s)

Bernhard Meindl

References

Templ, M. and Kowarik, A. and Meindl, B. Statistical Disclosure Control for Micro-Data Using the R Package sdcMicro. Journal of Statistical Software, 67 (4), 1–36, 2015. doi:10.18637/jss.v067.i04

Templ, M. Statistical Disclosure Control for Microdata: Methods and Applications in R. Springer International Publishing, 287 pages, 2017. ISBN 978-3-319-50272-4. doi:10.1007/978-3-319-50272-4 doi:10.1007/978-3-319-50272-4

Examples

## for objects of class sdcMicro:
data(testdata2)
testdata2$urbrur <- as.factor(testdata2$urbrur)
sdc <- createSdcObj(testdata2,
  keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
  numVars=c('expend','income','savings'), w='sampling_weight')
sdc <- groupAndRename(sdc, var="urbrur", before=c("1","2"), after=c("1"))

Description

Measures IL_correl() and IL_variables() were proposed by Andrzej Mlodak and are (theoretically) bounded between 0 and 1.

Usage

IL_correl(x, xm)

## S3 method for class 'il_correl'
print(x, digits = 3, ...)

IL_variables(x, xm)

## S3 method for class 'il_variables'
print(x, digits = 3, ...)

Arguments

Details

• IL_correl(): is a information-loss measure that can be applied to common numerically scaled variables in x and xm. It is based on diagonal entries of inverse correlation matrices in the original and perturbed data.

• IL_variables(): for common-variables in x and xm the individual distance-functions depend on the class of the variable; specifically these functions are different for numeric variables, ordered-factors and character/factor variables. The individual distances are summed up and scaled by n * m with n being the number of records and m being the number of (common) variables.

Details can be found in the references below

The implementation of IL_correl() differs slightly with the original proposition from Mlodak, A. (2020) as the constant multiplier was changed to 1 / sqrt(2) instead of 1/2 for better efficiency and interpretability of the measure.

Value

the corresponding information-loss measure

Author(s)

Bernhard Meindl [email protected]

References

Mlodak, A. (2020). Information loss resulting from statistical disclosure control of output data, Wiadomosci Statystyczne. The Polish Statistician, 2020, 65(9), 7-27, DOI: 10.5604/01.3001.0014.4121

Mlodak, A. (2019). Using the Complex Measure in an Assessment of the Information Loss Due to the Microdata Disclosure Control, Przegląd Statystyczny, 2019, 66(1), 7-26, DOI: 10.5604/01.3001.0013.8285

Examples

data("Tarragona", package = "sdcMicro")
res1 <- addNoise(obj = Tarragona, variables = colnames(Tarragona), noise = 100)
IL_correl(x = as.data.frame(res1$x), xm = as.data.frame(res1$xm))

res2 <- addNoise(obj = Tarragona, variables = colnames(Tarragona), noise = 25) 
IL_correl(x = as.data.frame(res2$x), xm = as.data.frame(res2$xm))

# creating test-inputs
n <- 150
x <- xm <- data.frame(
  v1 = factor(sample(letters[1:5], n, replace = TRUE), levels = letters[1:5]),
  v2 = rnorm(n),
  v3 = runif(3),
  v4 = ordered(sample(LETTERS[1:3], n, replace = TRUE), levels = c("A", "B", "C"))
)
xm$v1[1:5] <- "a"
xm$v2 <- rnorm(n, mean = 5)
xm$v4[1:5] <- "A"
IL_variables(x, xm)

Description

reads an sdcProblem with code that has been exported within sdcApp.

Usage

importProblem(path)

Arguments

Value

an object of class sdcMicro_GUI_export or an object of class 'simple.error'

Author(s)

Bernhard Meindl

Description

Estimation of the risk for each observation. After the risk is computed one can use e.g. the function localSuppr() for the protection of values of high risk. Further details can be found at the link given below.

Usage

indivRisk(x, method = "approx", qual = 1, survey = TRUE)

Arguments

Details

S4 class sdcMicro objects are only supported by function measure_risk that also estimates the individual risk with the same method.

Value

rk:

base individual risk

method:

method

qual:

final correction factor

fk:

frequency count

knames:

colnames of the key variables

Note

The base individual risk method was developed by Benedetti, Capobianchi and Franconi

Author(s)

Matthias Templ. Bug in method “exact” fixed since version 2.6.5. by Youri Baeyens.

References

Templ, M. and Kowarik, A. and Meindl, B. Statistical Disclosure Control for Micro-Data Using the R Package sdcMicro. Journal of Statistical Software, 67 (4), 1–36, 2015. doi:10.18637/jss.v067.i04

Franconi, L. and Polettini, S. (2004) Individual risk estimation in mu-Argus: a review. Privacy in Statistical Databases, Lecture Notes in Computer Science, 262–272. Springer

Machanavajjhala, A. and Kifer, D. and Gehrke, J. and Venkitasubramaniam, M. (2007) l-Diversity: Privacy Beyond k-Anonymity. ACM Trans. Knowl. Discov. Data, 1(1)

additionally, have a look at the vignettes of sdcMicro for further reading.

See Also

measure_risk, freqCalc

Examples

## example from Capobianchi, Polettini and Lucarelli:
data(francdat)
f <- freqCalc(francdat, keyVars=c(2,4,5,6),w=8)
f
f$fk
f$Fk
## individual risk calculation:
indivf <- indivRisk(f)
indivf$rk

Description

Calculate information loss after targeted record swapping using both the original and the swapped micro data. Information loss will be calculated on table counts defined by parameter 'table_vars' using either implemented information loss measures like absolute deviaton, relative absolute deviation and absolute deviation of square roots or custom metric, See details below.

Usage

infoLoss(
  data,
  data_swapped,
  table_vars,
  metric = c("absD", "relabsD", "abssqrtD"),
  custom_metric = NULL,
  hid = NULL,
  probs = sort(c(seq(0, 1, by = 0.1), 0.95, 0.99)),
  quantvals = c(0, 0.02, 0.05, 0.1, 0.2, 0.3, 0.4, 0.5, Inf),
  apply_quantvals = c("relabsD", "abssqrtD"),
  exclude_zeros = FALSE,
  only_inner_cells = FALSE
)

Arguments

Details

First frequency tables are build from both 'data' and 'data_swapped' using the variables defined in 'table_vars'. By default also all table margins will be calculated, see parameter 'only_inner_cells = FALSE'. After that the information loss metrices defined in either 'metric' or 'custom_metric' are applied on each of the table cells from both frequency tables. This is done in the sense of 'metric(x,y)' where 'metric' is the information loss, 'x' a cell from the table created from 'data' and 'y' the same cell from the table created from 'data_swapped'. One or more custom metrices can be applied using the parameter 'custom_metric', see also examples.

Value

Returns a list containing:

* 'cellvalues': 'data.table' showing in a long format for each table cell the frequency counts for 'data' ~ 'count_o' and 'data_swapped' ~ 'count_s'. * 'overview': 'data.table' containing the disribution of the 'noise' in number of cells and percentage. The 'noise' ist calculated as the difference between the cell values of the frequency table generated from the original and swapped data * 'measures': 'data.table' containing the quantiles and mean (column 'waht') of the distribution of the information loss metrices applied on each table cell. The quantiles are defined by parameter 'probs'. * 'cumdistr\*': 'data.table' containing the cumulative distribution of the information loss metrices. Distribution is shown in number of cells ('cnt') and percentage ('pct'). Column 'cat' shows all unique values of the information loss metric or the grouping defined by 'quantvals'. * 'false_zero': number of table cells which are non-zero when using 'data' and zero when using 'data_swapped'. * 'false_nonzero': number of table cells which are zero when using 'data' and non-zero when using 'data_swapped'. * 'exclude_zeros': value passed to 'exclude_zero' when calling the function.

Examples

# generate dummy data 
seed <- 2021
set.seed(seed)
nhid <- 10000
dat <- createDat( nhid )

# define paramters for swapping
k_anonymity <- 1
swaprate <- .05
similar <- list(c("hsize"))
hier <- c("nuts1","nuts2")
carry_along <- c("nuts3","lau2")
risk_variables <- c("ageGroup","national")
hid <- "hid"

# # apply record swapping
# dat_s <- recordSwap(data = dat, hid = hid, hierarchy = hier,
#                     similar = similar, swaprate = swaprate,
#                     k_anonymity = k_anonymity,
#                     risk_variables = risk_variables,
#                     carry_along = carry_along,
#                     return_swapped_id = TRUE,
#                     seed=seed)
# 
# 
# # calculate informationn loss
# # for the table nuts2 x national
# iloss <- infoLoss(data=dat, data_swapped = dat_s,
#                   table_vars = c("nuts2","national"))
# iloss$measures # distribution of information loss measures
# iloss$false_zero # no false zeros
# iloss$false_nonzero # no false non-zeros
# 
# # frequency tables of households accross
# # nuts2 x hincome
# 
# iloss <- infoLoss(data=dat, data_swapped = dat_s,
 #                  table_vars = c("nuts2","hincome"),
#                   hid = "hid")
# iloss$measures  
# 
# # define custom metric
# squareD <- function(x,y){
#   (x-y)^2
# }
# 
# iloss <- infoLoss(data=dat, data_swapped = dat_s,
#                  table_vars = c("nuts2","national"),
#                  custom_metric = list(squareD=squareD))
# iloss$measures # includes custom loss as well
#

Description

returns the number of observations violating k-anonymity.

Usage

kAnon_violations(object, weighted, k)

## S4 method for signature 'sdcMicroObj,logical,numeric'
kAnon_violations(object, weighted, k)

Arguments

Value

the number of records that are violating k-anonymity based on unweighted sample data only (in case parameter weighted is FALSE) or computing the number of observations that are estimated to violate k-anonymity in the population in case parameter weighted equals TRUE.

Description

To be used on both categorical and numeric input variables, although usage on categorical variables is the focus of the development of this software.

Usage

LocalRecProg(
  obj,
  ancestors = NULL,
  ancestor_setting = NULL,
  k_level = 2,
  FindLowestK = TRUE,
  weight = NULL,
  lowMemory = FALSE,
  missingValue = NA,
  ...
)

Arguments

Details

Each record in the data represents a category of the original data, and hence all records in the input data should be unique by the N Input Variables. To achieve bigger category sizes (k-anoymity), one can form new categories based on the recoding result and repeatedly apply this algorithm.

Value

dataframe with original variables and the supressed variables (suffix _lr). / the modified sdcMicroObj-class

Methods

list("signature(obj=\"sdcMicroObj\")")

Author(s)

Alexander Kowarik, Bernd Prantner, IHSN C++ source, Akimichi Takemura

References

Kowarik, A. and Templ, M. and Meindl, B. and Fonteneau, F. and Prantner, B.: Testing of IHSN Cpp Code and Inclusion of New Methods into sdcMicro, in: Lecture Notes in Computer Science, J. Domingo-Ferrer, I. Tinnirello (editors.); Springer, Berlin, 2012, ISBN: 978-3-642-33626-3, pp. 63-77. doi:10.1007/978-3-642-33627-0_6

Examples

data(testdata2)
cat_vars <- c("urbrur", "roof", "walls", "water", "sex", "relat")
anc_var <- c("water2", "water3", "relat2")
anc_setting <- c("water","water","relat")

r1 <- LocalRecProg(
  obj = testdata2,
  categorical = cat_vars,
  missingValue = -99)
r2 <- LocalRecProg(
  obj = testdata2,
  categorical = cat_vars,
  ancestor = anc_var,
  ancestor_setting = anc_setting,
  missingValue = -99)
r3 <- LocalRecProg(
  obj = testdata2,
  categorical = cat_vars,
  ancestor = anc_var,
  ancestor_setting = anc_setting,
  missingValue = -99,
  FindLowestK = FALSE)

# for objects of class sdcMicro:
sdc <- createSdcObj(
  dat = testdata2,
  keyVars = c("urbrur", "roof", "walls", "water", "electcon", "relat", "sex"),
  numVars = c("expend", "income", "savings"),
  w = "sampling_weight")
sdc <- LocalRecProg(sdc)

Description

A simple method to perfom local suppression.

Usage

localSupp(obj, threshold = 0.15, keyVar)

Arguments

Details

Values of high risk (above the threshold) of a certain variable (parameter keyVar) are suppressed.

Value

an updated object of class freqCalc or the sdcMicroObj-class object with manipulated data.

Author(s)

Matthias Templ and Bernhard Meindl

References

Templ, M. Statistical Disclosure Control for Microdata Using the R-Package sdcMicro, Transactions on Data Privacy, vol. 1, number 2, pp. 67-85, 2008. http://www.tdp.cat/issues/abs.a004a08.php

Templ, M. Statistical Disclosure Control for Microdata: Methods and Applications in R. Springer International Publishing, 287 pages, 2017. ISBN 978-3-319-50272-4. doi:10.1007/978-3-319-50272-4 doi:10.1007/978-3-319-50272-4

See Also

freqCalc, indivRisk

Examples

data(francdat)
keyVars <- paste0("Key",1:4)

f <- freqCalc(francdat, keyVars = keyVars, w = 8)
f
f$fk
f$Fk

## individual risk calculation:
indivf <- indivRisk(f)
indivf$rk

## Local Suppression
localS <- localSupp(f, keyVar = "Key4", threshold = 0.15)
f2 <- freqCalc(localS$freqCalc, keyVars = keyVars, w = 8)
indivf2 <- indivRisk(f2)
indivf2$rk
identical(indivf$rk, indivf2$rk)

## select another keyVar and run localSupp once again,
# if you think the table is not fully protected

## for objects of class sdcMicro:
data(testdata)
sdc <- createSdcObj(
  dat = testdata,
  keyVars = c("urbrur", "roof", "walls", "water", "electcon", "relat", "sex"),
  w = "sampling_weight"
)
sdc <- localSupp(sdc, keyVar = "urbrur", threshold = 0.045)
print(sdc, type = "ls")

Description

Algorithm to achieve k-anonymity by performing local suppression.

Usage

localSuppression(obj, k = 2, importance = NULL, combs = NULL, ...)

kAnon(obj, k = 2, importance = NULL, combs = NULL, ...)

Arguments

Details

The algorithm provides a k-anonymized data set by suppressing values in key variables. The algorithm tries to find an optimal solution to suppress as few values as possible and considers the specified importance vector. If not specified, the importance vector is constructed in a way such that key variables with a high number of characteristics are considered less important than key variables with a low number of characteristics.

The implementation provides k-anonymity per strata, if slot 'strataVar' has been set in sdcMicroObj-class or if parameter 'strataVar' is used when appying the data.frame method. For details, have a look at the examples provided.

Value

Manipulated data set with suppressions that has k-anonymity with respect to specified key-variables or the manipulated data stored in the sdcMicroObj-class.

Note

Deprecated methods 'localSupp2' and 'localSupp2Wrapper' are no longer available in sdcMicro > 4.5.0. kAnon is a more intutitive term for localSuppression because the aim is always to obtain k-anonymity for some parts of the data.

Author(s)

Bernhard Meindl, Matthias Templ

References

Templ, M. Statistical Disclosure Control for Microdata: Methods and Applications in R. Springer International Publishing, 287 pages, 2017. ISBN 978-3-319-50272-4. doi:10.1007/978-3-319-50272-4

Templ, M. and Kowarik, A. and Meindl, B. Statistical Disclosure Control for Micro-Data Using the R Package sdcMicro. Journal of Statistical Software, 67 (4), 1–36, 2015. doi:10.18637/jss.v067.i04

Examples

data(francdat)

## Local Suppression
localS <- localSuppression(francdat, keyVar=c(4,5,6))
localS
plot(localS)

## for objects of class sdcMicro, no stratification
data(testdata2)
kv <- c("urbrur", "roof", "walls", "water", "electcon", "relat", "sex")
sdc <- createSdcObj(testdata2, keyVars = kv, w = "sampling_weight")
sdc <- localSuppression(sdc)

## for objects of class sdcMicro, with stratification
testdata2$ageG <- cut(testdata2$age, 5, labels=paste0("AG",1:5))
sdc <- createSdcObj(
  dat = testdata2,
  keyVars = kv,
  w = "sampling_weight",
  strataVar = "ageG"
)
sdc <- localSuppression(sdc, nc = 1)

## it is also possible to provide k-anonymity for subsets of key-variables
## with different parameter k!
## in this case we want to provide 10-anonymity for all combinations
## of 5 key variables, 20-anonymity for all combinations with 4 key variables
## and 30-anonymity for all combinations of 3 key variables.
sdc <- createSdcObj(testdata2, keyVars = kv, w = "sampling_weight")
combs <- 5:3
k <- c(10, 20, 30)
sdc <- localSuppression(sdc, k = k, combs = combs)

## data.frame method (no stratification)
inp <- testdata2[,c(kv, "ageG")]
ls <- localSuppression(inp, keyVars = 1:7)
print(ls)
plot(ls)

## data.frame method (with stratification)
ls <- kAnon(inp, keyVars = 1:7, strataVars = 8)
print(ls)
plot(ls)

Description

Function to perform a fast and simple (primitive) method of microaggregation. (for large datasets)

Usage

mafast(obj, variables = NULL, by = NULL, aggr = 3, measure = mean)

Arguments

Value

If ‘obj’ was of class sdcMicroObj-class the corresponding slots are filled, like manipNumVars, risk and utility. If ‘obj’ was of class “data.frame” or “matrix” an object of the same class is returned.

Author(s)

Alexander Kowarik

See Also

microaggregation

Examples

data(Tarragona)
m1 <- mafast(Tarragona, variables=c("GROSS.PROFIT","OPERATING.PROFIT","SALES"),aggr=3)
data(testdata)
m2 <- mafast(testdata,variables=c("expend","income","savings"),aggr=50,by="sex")
summary(m2)

## for objects of class sdcMicro:
data(testdata2)
sdc <- createSdcObj(testdata2,
  keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
  numVars=c('expend','income','savings'), w='sampling_weight')
sdc <- dRisk(sdc)
sdc@risk$numeric
sdc1 <- mafast(sdc,aggr=4)
sdc1@risk$numeric

sdc2 <- mafast(sdc,aggr=10)
sdc2@risk$numeric

### Performance tests
x <- testdata
for(i in 1:20){
  x <- rbind(x,testdata)
}
system.time({
  xx <- mafast(
    obj = x,
    variables = c("expend", "income", "savings"),
    aggr = 50,
    by = "sex"
  )
})

Description

The function measures the disclosure risk for weighted or unweighted data. It computes the individual risk (and household risk if reasonable) and the global risk. It also computes a risk threshold based on a global risk value.

Prints a 'measure_risk'-object

Prints a 'ldiversity'-object

Usage

measure_risk(obj, ...)

ldiversity(obj, ldiv_index = NULL, l_recurs_c = 2, missing = -999, ...)

## S3 method for class 'measure_risk'
print(x, ...)

## S3 method for class 'ldiversity'
print(x, ...)

Arguments

Details

To be used when risk of disclosure for individuals within a family is considered to be statistical independent.

Internally, function freqCalc() and indivRisk are used for estimation.

Measuring individual risk: The individual risk approach based on so-called super-population models. In such models population frequency counts are modeled given a certain distribution. The estimation procedure of sample frequency counts given the population frequency counts is modeled by assuming a negative binomial distribution. This is used for the estimation of the individual risk. The extensive theory can be found in Skinner (1998), the approximation formulas for the individual risk used is described in Franconi and Polettini (2004).

Measuring hierarchical risk: If “hid” - the index of variable holding information on the hierarchical cluster structures (e.g., individuals that are clustered in households) - is provided, the hierarchical risk is additional estimated. Note that the risk of re-identifying an individual within a household may also affect the probability of disclosure of other members in the same household. Thus, the household or cluster-structure of the data must be taken into account when estimating disclosure risks. It is commonly assumed that the risk of re-identification of a household is the risk that at least one member of the household can be disclosed. Thus this probability can be simply estimated from individual risks as 1 minus the probability that no member of the household can be identified.

Global risk: The sum of the individual risks in the dataset gives the expected number of re-identifications that serves as measure of the global risk.

l-Diversity: If “ldiv_index” is unequal to NULL, i.e. if the indices of sensible variables are specified, various measures for l-diversity are calculated. l-diverstiy is an extension of the well-known k-anonymity approach where also the uniqueness in sensible variables for each pattern spanned by the key variables are evaluated.

Value

A modified sdcMicroObj-class object or a list with the following elements:

global_risk_ER:

expected number of re-identification.

global_risk:

global risk (sum of indivdual risks).

global_risk_pct:

global risk in percent.

Res:

matrix with the risk, frequency in the sample and grossed-up frequency in the population (and the hierachical risk) for each observation.

global_threshold:

for a given max_global_risk the threshold for the risk of observations.

max_global_risk:

the input max_global_risk of the function.

hier_risk_ER:

expected number of re-identification with household structure.

hier_risk:

global risk with household structure (sum of indivdual risks).

hier_risk_pct:

global risk with household structure in percent.

ldiverstiy:

Matrix with Distinct_Ldiversity, Entropy_Ldiversity and Recursive_Ldiversity for each sensitivity variable.

Prints risk-information into the console

Information on L-Diversity Measures in the console

Author(s)

Alexander Kowarik, Bernhard Meindl, Matthias Templ, Bernd Prantner, minor parts of IHSN C++ source

References

Franconi, L. and Polettini, S. (2004) Individual risk estimation in mu-Argus: a review. Privacy in Statistical Databases, Lecture Notes in Computer Science, 262–272. Springer

Machanavajjhala, A. and Kifer, D. and Gehrke, J. and Venkitasubramaniam, M. (2007) l-Diversity: Privacy Beyond k-Anonymity. ACM Trans. Knowl. Discov. Data, 1(1)

Templ, M. Statistical Disclosure Control for Microdata: Methods and Applications in R. Springer International Publishing, 287 pages, 2017. ISBN 978-3-319-50272-4. doi:10.1007/978-3-319-50272-4.

#' Templ, M. and Kowarik, A. and Meindl, B. Statistical Disclosure Control for Micro-Data Using the R Package sdcMicro. Journal of Statistical Software, 67 (4), 1–36, 2015. doi:10.18637/jss.v067.i04

See Also

freqCalc, indivRisk

measure_risk

Examples

## measure_risk with sdcMicro objects:
data(testdata)

sdc <- createSdcObj(testdata,
  keyVars=c('urbrur','roof','walls','water','electcon'),
numVars=c('expend','income','savings'), w='sampling_weight')

## risk is already estimated and available in...
names(sdc@risk)

## measure risk on data frames or matrices:
res <- measure_risk(testdata,
  keyVars=c("urbrur","roof","walls","water","sex"))
print(res)
head(res$Res)
resw <- measure_risk(testdata,
  keyVars=c("urbrur","roof","walls","water","sex"),w="sampling_weight")
print(resw)
head(resw$Res)
res1 <- ldiversity(testdata,
  keyVars=c("urbrur","roof","walls","water","sex"),ldiv_index="electcon")
print(res1)
head(res1)
res2 <- ldiversity(testdata,
  keyVars=c("urbrur","roof","walls","water","sex"),ldiv_index=c("electcon","relat"))
print(res2)
head(res2)

# measure risk with household risk
resh <- measure_risk(testdata,
  keyVars=c("urbrur","roof","walls","water","sex"),w="sampling_weight",hid="ori_hid")
print(resh)

# change max_global_risk
rest <- measure_risk(testdata,
  keyVars=c("urbrur","roof","walls","water","sex"),
  w="sampling_weight",max_global_risk=0.0001)
print(rest)

## for objects of class sdcMicro:
data(testdata2)
sdc <- createSdcObj(testdata2,
  keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
  numVars=c('expend','income','savings'), w='sampling_weight')
## -> when using `createSdcObj()`, the risks are already internally computed
## and it is not required to explicitely run `sdc <- measure_risk(sdc)`

Description

Replaces the raw household-level data with the anonymized household-level data in the full dataset for anonymization of data with a household structure (or other hierarchical structure). Requires a matching household ID in both files.

Usage

mergeHouseholdData(dat, hhId, dathh)

Arguments

Value

a data.frame with the treated household level variables and the raw individual level variables

Author(s)

Thijs Benschop and Bernhard Meindl

Examples

## Load data
x <- testdata

## donttest is necessary because of 
## Examples with CPU time > 2.5 times elapsed time
## caused by using C++ code and/or data.table
## Create household level dataset
x_hh <- selectHouseholdData(dat=x, hhId="ori_hid",
  hhVars=c("urbrur", "roof",  "walls", "water", "electcon", "household_weights"))
## Anonymize household level dataset and extract data
sdc_hh <- createSdcObj(x_hh, keyVars=c('urbrur','roof'), w='household_weights')
sdc_hh <- kAnon(sdc_hh, k = 3)
x_hh_anon <- extractManipData(sdc_hh)
 
## Merge anonymized household level data back into the full dataset
x_anonhh <- mergeHouseholdData(x, "ori_hid", x_hh_anon)
 
## Anonymize full dataset and extract data
sdc_full <- createSdcObj(x_anonhh, keyVars=c('sex', 'age', 'urbrur', 'roof'), w='sampling_weight')
sdc_full <- kAnon(sdc_full, k = 3)
x_full_anon <- extractManipData(sdc_full)

Description

Function to perform various methods of microaggregation.

Usage

microaggregation(
  obj,
  variables = NULL,
  aggr = 3,
  strata_variables = NULL,
  method = "mdav",
  weights = NULL,
  nc = 8,
  clustermethod = "clara",
  measure = "mean",
  trim = 0,
  varsort = 1,
  transf = "log"
)

Arguments

Details

On https://research.cbs.nl/casc/glossary.htm one can found the “official” definition of microaggregation:

Records are grouped based on a proximity measure of variables of interest, and the same small groups of records are used in calculating aggregates for those variables. The aggregates are released instead of the individual record values.

The recommended method is “rmd” which forms the proximity using multivariate distances based on robust methods. It is an extension of the well-known method “mdav”. However, when computational speed is important, method “mdav” is the preferable choice.

While for the proximity measure very different concepts can be used, the aggregation itself is naturally done with the arithmetic mean. Nevertheless, other measures of location can be used for aggregation, especially when the group size for aggregation has been taken higher than 3. Since the median seems to be unsuitable for microaggregation because of being highly robust, other mesures which are included can be chosen. If a complex sample survey is microaggregated, the corresponding sampling weights should be determined to either aggregate the values by the weighted arithmetic mean or the weighted median.

This function contains also a method with which the data can be clustered with a variety of different clustering algorithms. Clustering observations before applying microaggregation might be useful. Note, that the data are automatically standardised before clustering.

The usage of clustering method ‘Mclust’ requires package mclust02, which must be loaded first. The package is not loaded automatically, since the package is not under GPL but comes with a different licence.

The are also some projection methods for microaggregation included. The robust version ‘pppca’ or ‘clustpppca’ (clustering at first) are fast implementations and provide almost everytime the best results.

Univariate statistics are preserved best with the individual ranking method (we called them ‘onedims’, however, often this method is named ‘individual ranking’), but multivariate statistics are strong affected.

With method ‘simple’ one can apply microaggregation directly on the (unsorted) data. It is useful for the comparison with other methods as a benchmark, i.e. replies the question how much better is a sorting of the data before aggregation.

Value

If ‘obj’ was of class sdcMicroObj-class the corresponding slots are filled, like manipNumVars, risk and utility. If ‘obj’ was of class “data.frame”, an object of class “micro” with following entities is returned:

x:

original data

mx:

the microaggregated dataset

method:

method

aggr:

aggregation level

measure:

proximity measure for aggregation

Note

if only one variable is specified, mafast is applied and argument method is ignored. Parameters measure are ignored for methods mdav and rmd.

Author(s)

Matthias Templ, Bernhard Meindl

For method “mdav”: This work is being supported by the International Household Survey Network (IHSN) and funded by a DGF Grant provided by the World Bank to the PARIS21 Secretariat at the Organisation for Economic Co-operation and Development (OECD). This work builds on previous work which is elsewhere acknowledged.

Author for the integration of the code for mdav in R: Alexander Kowarik.

References

Templ, M. and Meindl, B., Robust Statistics Meets SDC: New Disclosure Risk Measures for Continuous Microdata Masking, Lecture Notes in Computer Science, Privacy in Statistical Databases, vol. 5262, pp. 113-126, 2008.

Templ, M. Statistical Disclosure Control for Microdata Using the R-Package sdcMicro, Transactions on Data Privacy, vol. 1, number 2, pp. 67-85, 2008. http://www.tdp.cat/issues/abs.a004a08.php

Templ, M. New Developments in Statistical Disclosure Control and Imputation: Robust Statistics Applied to Official Statistics, Suedwestdeutscher Verlag fuer Hochschulschriften, 2009, ISBN: 3838108280, 264 pages.

Templ, M. Statistical Disclosure Control for Microdata: Methods and Applications in R. Springer International Publishing, 287 pages, 2017. ISBN 978-3-319-50272-4. doi:10.1007/978-3-319-50272-4 doi:10.1007/978-3-319-50272-4

Templ, M. and Meindl, B. and Kowarik, A.: Statistical Disclosure Control for Micro-Data Using the R Package sdcMicro, Journal of Statistical Software, 67 (4), 1–36, 2015.

See Also

summary.micro, plotMicro, valTable

Examples

data(testdata)
# donttest since Examples with CPU time larger 2.5 times elapsed time, because
# of using data.table and multicore computation.

m <- microaggregation(
  obj = testdata[1:100, c("expend", "income", "savings")],
  method = "mdav",
  aggr = 4
)
summary(m)

## for objects of class sdcMicro:
## no stratification because `@strataVar` is `NULL`
data(testdata2)
sdc <- createSdcObj(
  dat = testdata2,
  keyVars = c("urbrur", "roof", "walls", "water", "electcon", "sex"),
  numVars = c("expend", "income", "savings"),
  w = "sampling_weight"
)
sdc <- microaggregation(
  obj = sdc,
  variables = c("expend", "income")
)

## with stratification using variable `"relat"`
strataVar(sdc) <- "relat"
sdc <- microaggregation(
  obj = sdc,
  variables = "savings"
)

Description

The microaggregation is based on the distances computed similar to the Gower distance. The distance function makes distinction between the variable types factor,ordered,numerical and mixed (semi-continuous variables with a fixed probability mass at a constant value e.g. 0)

Usage

microaggrGower(
  obj,
  variables = NULL,
  aggr = 3,
  dist_var = NULL,
  by = NULL,
  mixed = NULL,
  mixed.constant = NULL,
  trace = FALSE,
  weights = NULL,
  numFun = mean,
  catFun = VIM::sampleCat,
  addRandom = FALSE
)

Arguments

Details

The function sampleCat samples with probabilities corresponding to the occurrence of the level in the NNs. The function maxCat chooses the level with the most occurrences and random if the maximum is not unique.

Value

The function returns the updated sdcMicroObj or simply an altered data frame.

Note

In each by group all distance are computed, therefore introducing more by-groups significantly decreases the computation time and memory consumption.

Author(s)

Alexander Kowarik

See Also

sampleCat and maxCat

Examples

data(testdata,package="sdcMicro")
testdata <- testdata[1:200,]

for(i in c(1:7,9)) testdata[,i] <- as.factor(testdata[,i])
test <- microaggrGower(testdata,variables=c("relat","age","expend"),
  dist_var=c("age","sex","income","savings"),by=c("urbrur","roof"))

sdc <- createSdcObj(testdata,
  keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
  numVars=c('expend','income','savings'), w='sampling_weight')

sdc <- microaggrGower(sdc)

Description

Small aritificial toy data set.

Format

The format is: num [1:13, 1:5] 5 7 2 1 7 8 12 3 15 4 ... - attr(*, "dimnames")=List of 2 ..$ : chr [1:13] "10000" "11000" "12000" "12100" ... ..$ : chr [1:5] "one" "two" "three" "four" ...

Examples

data(microData)
microData <- as.data.frame(microData)
m1 <- microaggregation(microData, method="mdav")
summary(m1)

Description

The sample frequencies are assumed to be independent and following a Poisson distribution. The parameters of the corresponding parameters are estimated by a log-linear model including the main effects and possible interactions.

Usage

modRisk(obj, method = "default", weights, formulaM, bound = Inf, ...)

Arguments

Details

This measure aims to (1) calculate the number of sample uniques that are population uniques with a probabilistic Poisson model and (2) to estimate the expected number of correct matches for sample uniques.

ad 1) this risk measure is defined over all sample uniques as

$\tau_1 = \sum\limits_{j:f_j=1} P(F_j=1 | f_j=1) \quad ,$

i.e. the expected number of sample uniques that are population uniques.

ad 2) this risk measure is defined over all sample uniques as

$\tau_2 = \sum\limits_{j:f_j=1} P(1 / F_j | f_j=1) \quad .$

Since population frequencies $F_k$ are unknown, they need to be estimated.

The iterative proportional fitting method is used to fit the parameters of the Poisson distributed frequency counts related to the model specified to fit the frequency counts. The obtained parameters are used to estimate a global risk, defined in Skinner and Holmes (1998).

Value

Two global risk measures and some model output given the specified model. If this method is applied to an sdcMicroObj-class-object, the slot 'risk' in the object ist updated with the result of the model-based risk-calculation.

Author(s)

Matthias Templ, Marius Totter, Bernhard Meindl

References

Skinner, C.J. and Holmes, D.J. (1998) Estimating the re-identification risk per record in microdata. Journal of Official Statistics, 14:361-372, 1998.

Rinott, Y. and Shlomo, N. (1998). A Generalized Negative Binomial Smoothing Model for Sample Disclosure Risk Estimation. Privacy in Statistical Databases. Lecture Notes in Computer Science. Springer-Verlag, 82–93.

Clogg, C.C. and Eliasson, S.R. (1987). Some Common Problems in Log-Linear Analysis. Sociological Methods and Research, 8-44.

See Also

loglm, measure_risk

Examples

## data.frame method
data(testdata2)
form <- ~sex+water+roof
w <- "sampling_weight"

(modRisk(testdata2, method = "default", formulaM = form, weights = w))
(modRisk(testdata2, method = "CE", formulaM = form, weights = w))
(modRisk(testdata2, method = "PML", formulaM = form, weights = w))
(modRisk(testdata2, method = "weightedLLM", formulaM = form, weights = w))
(modRisk(testdata2, method = "IPF", formulaM = form, weights = w))

## application to a sdcMicroObj
data(testdata2)
sdc <- createSdcObj(testdata2,
 keyVars = c("urbrur", "roof", "walls", "electcon", "relat", "sex"),
 numVars = c("expend", "income", "savings"),
 w = "sampling_weight")
sdc <- modRisk(sdc, form = ~sex+water+roof)
slot(sdc, "risk")$model



# an example using data from the laeken-pkg
library(laeken)
data(eusilc)
f <- as.formula(paste(" ~ ", "db040 + hsize + rb090 +
             age + pb220a + age:rb090 + age:hsize +
             hsize:rb090"))
w <- "rb050"
(modRisk(eusilc, method = "default", weights = w, formulaM = f, bound = 5))
(modRisk(eusilc, method = "CE", weights =  w, formulaM = f, bound = 5))
(modRisk(eusilc, method = "PML", weights = w, formulaM = f, bound = 5))
(modRisk(eusilc, method = "weightedLLM", weights = w, formulaM = f, bound = 5))

Description

Imputation and detection of outliers

Usage

mvTopCoding(x, maha = NULL, center = NULL, cov = NULL, alpha = 0.025)

Arguments

Details

Winsorizes the potential outliers on the ellipsoid defined by (robust) Mahalanobis distances in direction to the center of the data

Value

the imputed winsorized data

Author(s)

Johannes Gussenbauer, Matthias Templ

Examples

set.seed(123)
x <- MASS::mvrnorm(20, mu = c(5,5), Sigma = matrix(c(1,0.9,0.9,1), ncol = 2))
x[1, 1] <- 3
x[1, 2] <- 6
plot(x)
ximp <- mvTopCoding(x)
points(ximp, col = "blue", pch = 4)

# more dimensions
Sigma <- diag(5)
Sigma[upper.tri(Sigma)] <- 0.9
Sigma[lower.tri(Sigma)] <- 0.9
x <- MASS::mvrnorm(20, mu = rep(5,5), Sigma = Sigma)
x[1, 1] <- 3
x[1, 2] <- 6
pairs(x)

ximp <- mvTopCoding(x)
xnew <- data.frame(rbind(x, ximp))
xnew$beforeafter <- rep(c(0,1), each = nrow(x))
pairs(xnew, col = xnew$beforeafter, pch = 4)

# by hand (non-robust)
x[2,2] <- NA
m <- colMeans(x, na.rm = TRUE)
s <- cov(x, use = "complete.obs")
md <- stats::mahalanobis(x, m, s)
ximp <- mvTopCoding(x, center = m, cov = s, maha = md)
plot(x)
points(ximp, col = "blue", pch = 4)

Description

internal function used to provide the undo-functionality.

Usage

nextSdcObj(obj)

Arguments

Value

a modified sdcMicroObj-class object

Description

Reorders the data according to a column in the data set.
NOTE: This is an internal function used for testing the C++-function orderData which is used inside the C++-function recordSwap() to speed up performance.

Usage

orderData_cpp(data, orderIndex)

Arguments

Value

ordered data set.

Description

This function creates barplots to display the number of suppressed values in categorical key variables to achieve k-anonymity.

Usage

## S3 method for class 'localSuppression'
plot(x, ...)

Arguments

Value

a ggplot plot object

Author(s)

Bernhard Meindl, Matthias Templ

See Also

localSuppression()

Examples

data(francdat)

Plotfunctions for objects of class sdcMicroObj

Description

Descriptive plot function for sdcMicroObj-objects. Currently only visualization of local supression is implemented.

Usage

## S3 method for class 'sdcMicroObj'
plot(x, type = "ls", ...)

Arguments

Value

a ggplot plot object or (invisible) NULL if local suppression using kAnon() has not been applied

Author(s)

Bernhard Meindl

Examples

data(testdata)
sdc <- createSdcObj(testdata,
  keyVars = c("urbrur", "roof", "walls", "relat", "sex"),
  w = "sampling_weight")
sdc <- kAnon(sdc, k = 3)
plot(sdc, type = "ls")

Description

Plots for the comparison of the original data and perturbed data.

Usage

plotMicro(x, p, which.plot = 1:3)

Arguments

Details

Univariate and multivariate comparison plots are implemented to detect differences between the perturbed and the original data, but also to compare perturbed data which are produced by different methods.

Value

returns NULL; the selected plot is displayed

Author(s)

Matthias Templ

References

Templ, M. and Meindl, B., Software Development for SDC in R, Lecture Notes in Computer Science, Privacy in Statistical Databases, vol. 4302, pp. 347-359, 2006.

See Also

microaggregation()

Examples

data(free1)
df <- as.data.frame(free1)[, 31:34]
m1 <- microaggregation(df, method = "onedims", aggr = 3)
plotMicro(m1, p = 1, which.plot = 1)
plotMicro(m1, p = 1, which.plot = 2)
plotMicro(m1, p = 1, which.plot = 3)

Description

To be used on categorical data stored as factors. The algorithm randomly changes the values of variables in selected records (usually the risky ones) according to an invariant probability transition matrix or a custom-defined transition matrix.

Usage

pram(obj, variables = NULL, strata_variables = NULL, pd = 0.8, alpha = 0.5)

Arguments

Value

a modified sdcMicroObj object or a new object containing original and post-randomized variables (with suffix "_pram").

Note

Deprecated method 'pram_strata' is no longer available in sdcMicro > 4.5.0

Author(s)

Alexander Kowarik, Matthias Templ, Bernhard Meindl

References

https://www.gnu.org/software/glpk/

Kowarik, A. and Templ, M. and Meindl, B. and Fonteneau, F. and Prantner, B.: Testing of IHSN Cpp Code and Inclusion of New Methods into sdcMicro, in: Lecture Notes in Computer Science, J. Domingo-Ferrer, I. Tinnirello (editors.); Springer, Berlin, 2012, ISBN: 978-3-642-33626-3, pp. 63-77. doi:10.1007/978-3-642-33627-0_6

Templ, M. and Kowarik, A. and Meindl, B.: Statistical Disclosure Control for Micro-Data Using the R Package sdcMicro. in: Journal of Statistical Software, 67 (4), 1–36, 2015. doi:10.18637/jss.v067.i04

Templ, M.: Statistical Disclosure Control for Microdata: Methods and Applications in R. in: Springer International Publishing, 287 pages, 2017. ISBN 978-3-319-50272-4. doi:10.1007/978-3-319-50272-4

Examples

data(testdata)

## donttest is necessary because of
## Examples with CPU time > 2.5 times elapsed time
## caused by using C++ code and/or data.table
## using a factor variable as input
res <- pram(as.factor(testdata$roof))
print(res)
summary(res)

## using a data.frame as input
## pram can only be applied to factors
## -- > we have to recode to factors beforehand
testdata$roof <- factor(testdata$roof)
testdata$walls <- factor(testdata$walls)
testdata$water <- factor(testdata$water)

## pram() is applied within subgroups defined by
## variables "urbrur" and "sex"
res <- pram(
  obj = testdata,
  variables = "roof",
 strata_variables = c("urbrur", "sex"))
print(res)
summary(res)

## default parameters (pd = 0.8 and alpha = 0.5) for the generation
## of the invariant transition matrix will be used for all variables
res1 <- pram(
  obj = testdata,
  variables = c("roof", "walls", "water"))
print(res1)

## specific parameter settings for each variable
res2 <- pram(
   obj = testdata,
   variables = c("roof", "walls", "water"),
   pd = c(0.95, 0.8, 0.9),
   alpha = 0.5)
print(res2)

## detailed information on pram-parameters (such as the transition matrix 'Rs')
## is stored in the output, eg. for variable 'roof'
#attr(res2, "pram_params")$roof
## we can also specify a custom transition-matrix directly
mat <- diag(length(levels(testdata$roof)))
rownames(mat) <- colnames(mat) <- levels(testdata$roof)
res3 <- pram(
   obj = testdata,
   variables = "roof",
  pd = mat)
print(res3) # of course, nothing has changed!
## it is possible use a transition matrix for a variable and use the 'traditional' way
## of specifying a number for the minimal diagonal entries of the transision matrix
## for other variables. In this case we must supply `pd` as list.
res4 <- pram(
  obj = testdata,
  variables = c("roof", "walls"),
  pd = list(mat, 0.5),
  alpha = c(NA, 0.5))
print(res4)
summary(res4)
attr(res4, "pram_params")

## application to objects of class sdcMicro with default parameters
data(testdata2)
testdata2$urbrur <- factor(testdata2$urbrur)
sdc <- createSdcObj(
  dat = testdata2,
  keyVars = c("roof", "walls", "water", "electcon", "relat", "sex"),
  numVars = c("expend", "income", "savings"),
  w = "sampling_weight")
sdc <- pram(
  obj = sdc,
  variables = "urbrur")
print(sdc, type = "pram")

## this is equal to the previous application. If argument 'variables' is NULL,
## all variables from slot 'pramVars' will be used if possible.
sdc <- createSdcObj(
   dat = testdata2,
  keyVars = c("roof", "walls", "water", "electcon", "relat", "sex"),
  numVars = c("expend", "income", "savings"),
  w = "sampling_weight",
  pramVars = "urbrur")
sdc <- pram(sdc)
print(sdc, type="pram")

## we can specify transition matrices for sdcMicroObj-objects too
testdata2$roof <- factor(testdata2$roof)
sdc <- createSdcObj(
  dat = testdata2,
  keyVars = c("roof", "walls", "water", "electcon", "relat", "sex"),
  numVars = c("expend", "income", "savings"),
  w = "sampling_weight")
mat <- diag(length(levels(testdata2$roof)))

rownames(mat) <- colnames(mat) <- levels(testdata2$roof)
mat[1,] <- c(0.9, 0, 0, 0.05, 0.05)
sdc <- pram(
   obj = sdc,
   variables = "roof",
   pd = mat)
print(sdc, type = "pram")

## we can also have a look at the transitions
get.sdcMicroObj(sdc, "pram")$transitions

Description

Print method for objects from class freqCalc.

Usage

## S3 method for class 'freqCalc'
print(x, ...)

Arguments

Value

information about the frequency counts for key variables for object of class freqCalc.

Author(s)

Matthias Templ

See Also

freqCalc

Examples

## example from Capobianchi, Polettini and Lucarelli:
data(francdat)
f <- freqCalc(francdat, keyVars=c(2,4,5,6),w=8)
f

Description

Print method for objects from class indivRisk

Usage

## S3 method for class 'indivRisk'
print(x, ...)

Arguments

Value

few information about the method and the final correction factor for objects of class ‘indivRisk’.

Author(s)

Matthias Templ

See Also

indivRisk

Examples

## example from Capobianchi, Polettini and Lucarelli:
data(francdat)
f1 <- freqCalc(francdat, keyVars=c(2,4,5,6),w=8)
data.frame(fk=f1$fk, Fk=f1$Fk)
## individual risk calculation:
indivRisk(f1)

Description

Print method for objects from class localSuppression

Usage

## S3 method for class 'localSuppression'
print(x, ...)

Arguments

Value

Information about the frequency counts for key variables for object of class ‘localSuppression’.

Author(s)

Matthias Templ

See Also

localSuppression

Examples

## example from Capobianchi, Polettini and Lucarelli:
data(francdat)
l1 <- localSuppression(francdat, keyVars=c(2,4,5,6))
l1

Description

printing an object of class micro

Usage

## S3 method for class 'micro'
print(x, ...)

Arguments

Value

information about method and aggregation level from objects of class micro.

Author(s)

Matthias Templ

See Also

microaggregation

Examples

data(free1)
free1 <- as.data.frame(free1)
m1 <- microaggregation(free1[, 31:34], method='onedims', aggr=3)
m1

Description

Print method for objects from class modrisk

Usage

## S3 method for class 'modrisk'
print(x, ...)

Arguments

Value

Output of model-based risk estimation

Author(s)

Bernhard Meindl

See Also

modRisk

Description

Print method for objects from class pram

Usage

## S3 method for class 'pram'
print(x, ...)

Arguments

Value

absolute and relative frequencies of changed observations in each modified variable

Author(s)

Bernhard Meindl, Matthias Templ

Matthias Templ and Bernhard Meindl

See Also

pram

Print and Extractor Functions for objects of class sdcMicroObj-class

Description

Descriptive print function for Frequencies, local Supression, Recoding, categorical risk and numerical risk.

Usage

## S4 method for signature 'sdcMicroObj'
print(x, type = "kAnon", docat = TRUE, ...)

Arguments

Details

Possible values for the type argument of the print function are: "freq": for Frequencies, "ls": for Local Supression output, "pram": for results of post-randomization "recode":for Recodes, "risk": forCategorical risk and "numrisk": for Numerical risk.

Possible values for the type argument of the freq function are: "fk": Sample frequencies and "Fk": weighted frequencies.

Author(s)

Alexander Kowarik, Matthias Templ, Bernhard Meindl

Examples

data(testdata)

sdc <- createSdcObj(testdata,
  keyVars=c('urbrur','roof','walls','relat','sex'),
  pramVars=c('water','electcon'),
  numVars=c('expend','income','savings'), w='sampling_weight')
sdc <- microaggregation(sdc, method="mdav", aggr=3)
print(sdc)
print(sdc, type="general")
print(sdc, type="ls")
print(sdc, type="recode")
print(sdc, type="risk")
print(sdc, type="numrisk")
print(sdc, type="pram")
print(sdc, type="kAnon")
print(sdc, type="comp_numvars")

Description

Print method for objects from class suda2.

Usage

## S3 method for class 'suda2'
print(x, ...)

Arguments

Value

Table of dis suda scores.

Author(s)

Matthias Templ

See Also

suda2

Description

Randomly select records given a probability weight vector prob.
NOTE: This is an internal function used for testing the C++-function randSample which is used inside the C++-function recordSwap().

Usage

randSample_cpp(ID, N, prob, IDused, seed)

Arguments

Description

Swapping values within a range so that, first, the correlation structure of original variables are preserved, and second, the values in each record are disturbed. To be used on numeric or ordinal variables where the rank can be determined and the correlation coefficient makes sense.

Usage

rankSwap(
  obj,
  variables = NULL,
  TopPercent = 5,
  BottomPercent = 5,
  K0 = NULL,
  R0 = NULL,
  P = NULL,
  missing = NA,
  seed = NULL
)

Arguments

Details

Rank swapping sorts the values of one numeric variable by their numerical values (ranking). The restricted range is determined by the rank of two swapped values, which cannot differ, by definition, by more than P percent of the total number of observations. Only positive P, R0 and K0 are used and only one of it must be supplied. If none is supplied, sdcMicro sets parameter r0 to 0.95 internally.

Value

The rank-swapped data set or a modified sdcMicroObj-class object.

Author(s)

Alexander Kowarik for the interface, Bernhard Meindl for improvements.

For the underlying C++ code: This work is being supported by the International Household Survey Network (IHSN) and funded by a DGF Grant provided by the World Bank to the PARIS21 Secretariat at the Organisation for Economic Co-operation and Development (OECD). This work builds on previous work which is elsewhere acknowledged.

References

Moore, Jr.R. (1996) Controlled data-swapping techniques for masking public use microdata, U.S. Bureau of the Census Statistical Research Division Report Series, RR 96-04.

Kowarik, A. and Templ, M. and Meindl, B. and Fonteneau, F. and Prantner, B.: Testing of IHSN Cpp Code and Inclusion of New Methods into sdcMicro, in: Lecture Notes in Computer Science, J. Domingo-Ferrer, I. Tinnirello (editors.); Springer, Berlin, 2012, ISBN: 978-3-642-33626-3, pp. 63-77. doi:10.1007/978-3-642-33627-0_6

Examples

data(testdata2)
data_swap <- rankSwap(
  obj = testdata2,
  variables = c("age", "income", "expend", "savings")
)

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