Movatterモバイル変換

tidydp 0.1.0

Initial CRAN Release

This is the first release of tidydp, a tidy-style interface forapplying differential privacy to data frames in R.

Core Features

• Differential Privacy Mechanisms
  • Laplace mechanism for pure epsilon-differential privacy
  • Gaussian mechanism for (epsilon, delta)-differential privacy
  • Automatic sensitivity calculations based on data bounds
• Tidy-Style API Functions
  • dp_add_noise(): Add calibrated noise to numeric columnswith pipe support
  • dp_count(): Compute differentially private counts withoptional grouping
  • dp_mean(): Compute differentially private means withoptional grouping
  • dp_sum(): Compute differentially private sums withoptional grouping
• Privacy Budget Management
  • new_privacy_budget(): Create and initialize privacybudgets
  • check_privacy_budget(): Verify sufficient budget beforeoperations
  • Automatic budget tracking with basic composition
  • Print method for budget status visualization

Technical Details

• All functions support the magrittr pipe operator(%>%)
• Flexible mechanism selection (Laplace or Gaussian)
• Support for grouped operations usinggroup_byparameter
• Comprehensive error handling and input validation
• Built from scratch without external differential privacydependencies

Documentation

• Complete function documentation with examples
• Comprehensive README with usage examples
• Full test suite with >95% code coverage
• Example script demonstrating common workflows

Notes

• This package implements differential privacy mechanisms from firstprinciples
• Suitable for statistical analysis with formal privacyguarantees
• Compatible with the tidyverse ecosystem
• Designed for CRAN submission standards