ldpdist: Local Differential Privacy Tools for Distributional Inference
Estimates cumulative distribution functions and quantiles under
local differential privacy. Provides binary and categorical randomization
mechanisms, isotonic distribution estimators, and simulation interfaces
for sequential quantile estimation with self-normalized inference and
asymptotic confidence sequences. Distribution estimators operate on
privatized reports; quantile interfaces simulate local randomization
internally. Methods include the quantile procedure of Liu et al. (2023)
<https://proceedings.mlr.press/v202/liu23q.html> and the distribution
estimator of Liu, Hu and Kong (2024)
<https://proceedings.mlr.press/v235/liu24z.html>.
Documentation:
Downloads:
Linking:
Please use the canonical form
https://CRAN.R-project.org/package=ldpdist
to link to this page.