Laplace mechanism (primary entry point)

Implements M(D) = f(D) + Lap(sensitivity / epsilon), Dwork-Roth Definition 3.3.

Usage

morie_laplc(value, sensitivity, epsilon, seed = NULL)

Arguments

value

Numeric scalar. The true query result f(D).

sensitivity

Numeric scalar. The L1 sensitivity of the query.

epsilon

Numeric scalar. The privacy budget (must be > 0).

seed

Integer or NULL. Optional seed for reproducibility.

Value

A named list (RichResult) with the noisy value and metadata.

Examples

morie_laplc(value = c(1, 2, 3, 4, 5, 6, 7, 8),
  sensitivity = c(1, 2, 3, 4, 5, 6, 7, 8), epsilon = c(1, 2, 3, 4, 5, 6, 7, 8))
#> $value
#> [1] -2.4521822 -1.4521822 -0.4521822  0.5478178  1.5478178  2.5478178  3.5478178
#> [8]  4.5478178
#> 
#> $true_value
#> [1] 1 2 3 4 5 6 7 8
#> 
#> $noise
#> [1] -3.452182
#> 
#> $scale
#> [1] 1 1 1 1 1 1 1 1
#> 
#> $sensitivity
#> [1] 1 2 3 4 5 6 7 8
#> 
#> $epsilon
#> [1] 1 2 3 4 5 6 7 8
#> 
#> $mechanism
#> [1] "laplace"
#>