rebalancedNoise: Perturbative Statistical Disclosure Control for Magnitude Tables

rebalancedNoise implements the EZS method for magnitude tables as described in the Paper Using Perturbative Methods for Magnitude Tables in Statistical Disclosure Control from Sabolová et al. (2025). It provides a high-performance framework for applying record-level noise with a dynamic rebalancing algorithm that preserves data quality in non-sensitive cells while ensuring additive consistency across complex hierarchies.

Key Features

• Hierarchy Integrity: Ensures that all hierarchical aggregates remain perfectly additive after perturbation.
• Dynamic Rebalancing: Minimizes noise in non-sensitive cells by "balancing" noise directions within hierarchical nodes.
• High Performance: Rebalancing algorithm is implemented in C++ with OpenMP support for fast processing of large inputs.
• Built-in Diagnostics: Powerful $summarize() tools to evaluate rebalancing efficiency and data utility via high-density statistical reports.
• Seamless Integration: Designed to work directly with data.table and sdcHierarchies.

Installation

# Install from Github
devtools::install_github("sdcTools/rebalancedNoise")

Quick Start

library(rebalancedNoise)

# Generate Dummy-Data
N <- 100
countries <- c("AT", "DE", "NL", "SE", "FR", "IT")
set.seed(1)
dt <- data.table(
  country = sample(countries, N, replace = TRUE),
  turnover = runif(N, 10, 1000),
  direction = sample(c(1, -1), N, replace = TRUE),
  noise_multiplier = 0.05
)

# Define hierarchy
dims <- list(
  country = sdcHierarchies::hier_create("Total", nodes = countries)
)

# Initialize the Object
sdc <- rn_setup(
    data = dt, 
    dim_list = dims, 
    num_vars = c("turnover"),
    sensitive_params = list(n_threshold = 5)
)

# Perturb the numerical variable
sdc$perturb("turnover")

# Show summarize-statistics (initial vs. final perturbation)
sdc$summarize("turnover")

# extract results
sdc$get_results("turnover", format = "long") # or "wide"

Documentation

For a detailed walkthrough of the method, multi-variable handling, and parallelization settings (via options() or environment variables), please see the full package vignette:

vignette("getting-started", package = "rebalancedNoise")