GaussSuppression (version 1.2.0)

IntervalLimits: Construct feasibility interval requirements

Description

Creates one or more sets of requirements for feasibility intervals of cell values in protected tables. These can include lower bound requirements ("lomax_"), upper bound requirements ("upmin_"), and/or minimum width requirements ("rlim_*"). The requirements may be specified as relative percentages or as absolute distances, and can be set symmetrically around the true cell value or separately for lower and upper bounds.

Usage

IntervalLimits(
  ...,
  protectionPercent = 0,
  protectionLimit = 0,
  loProtectionPercent = protectionPercent,
  loProtectionLimit = protectionLimit,
  rangePercent = 0,
  rangeMin = 0,
  primary,
  num,
  freq,
  freqVar,
  dominanceVar = NULL,
  intervalVar = NULL
)

Value

A matrix with column names constructed from the type of requirement and the associated variable name.

Arguments

...

Unused parameters

protectionPercent

Required distance in percent between the true cell value and the upper bound of the feasibility interval. If loProtectionPercent is not set, the same percentage distance is also required to the lower bound.

protectionLimit

Required absolute distance between the true cell value and the upper bound of the feasibility interval. If loProtectionLimit is not set, the same absolute distance is also required to the lower bound.

loProtectionPercent

Required distance in percent between the true cell value and the lower bound of the feasibility interval. Is by default set to protectionPercent, i.e. symmetrical requirements.

loProtectionLimit

Required absolute distance between the true cell value and the lower bound of the feasibility interval. Is by default set to protectionLimit, i.e. symmetrical requirements.

rangePercent

Minimum required width of the feasibility interval expressed as a percentage of the true cell value.

rangeMin

Minimum required width of the feasibility interval expressed as an absolute value.

primary

The primary vector generated by parent function

num

The num data frame generated by parent function

freq

The freq vector generated by parent function

freqVar

As input to e.g. SuppressSmallCounts()

dominanceVar

As input to e.g. SuppressDominantCells()

intervalVar

Numerical variable(s) used for interval calculations. If NULL, the variable names may be inferred from the num data frame as a consequence of interval requirements created by the primary function(s). If no variables can be inferred this way, dominanceVar, first numVar or freqVar will be used.

Examples

dat <- SSBtoolsData("magnitude1")
dat["num2"] <- 1:nrow(dat)

SuppressDominantCells(data = dat, 
    numVar = "value", 
    formula = ~sector2 * geo + sector4 * eu, 
    contributorVar = "company", 
    k = c(80, 99), 
    rangePercent = 10, rangeMin = 1,
    protectionPercent = 3, 
    protectionLimit = 5, loProtectionLimit = 4)

SuppressDominantCells(data = dat, 
    dominanceVar = "value",
    numVar = "num2", 
    formula = ~sector2 * geo + sector4 * eu, 
    contributorVar = "company", 
    pPercent = 10,
    intervalVar = c("value","freq", "num2"), 
    rangePercent = c(10, 10, 30), rangeMin = c(1, 0.2222, 2.222))
    

    
## Below are two alternative ways of calculating interval requirements. 
## In these cases, the requirements are generated by the primary suppression 
## functions when the parameter `protectionIntervals = TRUE` is specified.


# See ?MagnitudeRule
SuppressDominantCells(data = dat, 
                      dominanceVar = "value",
                      formula = ~sector2 * geo + sector4 * eu, 
                      contributorVar = "company", 
                      pPercent = 10,
                      protectionIntervals = TRUE)

# See ?PrimaryDefault
SuppressSmallCounts(data = dat, 
                    formula = ~sector2 * geo + sector4 * eu, 
                    maxN = 3,
                    protectionIntervals = TRUE)     
                    
                    
                    
## Combining IntervalLimits arguments 
## with protectionIntervals = TRUE also works

SuppressSmallCounts(data = dat, 
                    formula = ~sector2 * geo + sector4 * eu, 
                    maxN = 3,
                    protectionIntervals = TRUE,
                    protectionPercent = 50)