l2sens: Calculate the l2 sensitivity

In difuture-lmu/dsROCGLM: ROC and calibration analysis for DataSHIELD

Calculate the l2 sensitivity

Description

Calculation of the l2 sensitivity using a histogram representetation. Source: https://www.cis.upenn.edu/~aaroth/Papers/privacybook.pdf

Usage

l2sens(
  dat_name,
  scores_name,
  nbreaks = NULL,
  col_names = NULL,
  norm = diff,
  drop_on_error = TRUE
)

Arguments

dat_name	('character(1)') Name of the data used to find adjacent inputs.
scores_name	('character(1)') Name of the predicted scores/probabilities.
nbreaks	('integer(1L)') Number of breaks used for the histogram (default = nrow(dat) / 3).
col_names	('character(1)') Subset of columns used to find adjacent inputs.
norm	('function()') Function to calculate the differnece between the scores of adjacent inputs (default = Euclidean norm).
drop_on_error	('logical(1L)') NA is returned if an error occurs on a server.

Value

List with maximal l2 sensitivity, indices of inputs that are used to calculate the maximal l2 sensitivity, and the number of adjacent inputs.

Author(s)

Daniel S.

difuture-lmu/dsROCGLM documentation built on March 24, 2024, 1:07 p.m.