Calculate empirical frequencies from aggregated table

Calculates empirical frequencies of counts from an aggregated table constructed with tabulate_cnt_micro_data, apply_ckm, or tabulate_and_apply_ckm functions.

Usage

compute_frequencies(tableau, cat_vars, hrc_vars = NULL, cnt_var = "nb_obs")

Arguments

tableau

data.frame. Table generated with tabulate_cnt_micro_data, apply_ckm, or tabulate_and_apply_ckm functions

cat_vars

Character vector. Categorical variables

hrc_vars

Named list. Hierarchical variables

cnt_var

character vector indicating the name of the count variable

Value

data.frame with 3 columns:

• i: count value

• N: number of occurrences of the count

• p_hat: empirical frequency of the count

Details

The function estimates the number of zeroes from the structure of the input table. In some circumstances, the zeroes can be underestimated.

References

Enderle T, Giessing S, Tent R (2020). “Calculation of Risk Probabilities for the Cell Key Method.” Privacy in Statistical Databases, 12276, 151–165. doi:10.1007/978-3-030-57521-2_11 . http://link.springer.com/10.1007/978-3-030-57521-2_11. Jamme J (2025). “A Framework for Cell Key Method Parameters Calibration based on a Risk-Utility trade-off.” In Expert Meeting on Statistical Data Confidentiality. https://www.researchgate.net/publication/405051057_A_Framework_for_Cell_Key_Method_Parameters_Calibration_based_on_a_Risk-Utility_trade-off.

Examples

if (FALSE) { # \dontrun{
library(dplyr)
data("dtest")

cat_vars1 = c("DEP", "DIPLOME", "SEXE", "AGE")
hrc_vars1 = NULL
tab_comptage1 <- tabulate_cnt_micro_data(
  df = dtest, rk = NULL,
  cat_vars = cat_vars1,
  marge_label = "Total"
)
p_hat1 <- compute_frequencies(tab_comptage1, cat_vars1, hrc_vars1)

# With hierarchical variables:
cat_vars2 = c("DIPLOME", "SEXE", "AGE")
hrc_vars2 = list(GEO = c("REG","DEP"), TYPES = c("TYPE","TYPE2"))
tab_comptage2 <- tabulate_cnt_micro_data(
  df = dtest, rk = NULL,
  cat_vars = cat_vars2,
  hrc_vars = hrc_vars2,
  marge_label = "Total"
)
p_hat2 <- compute_frequencies(tab_comptage2, cat_vars2, hrc_vars2)
} # }