Local Differential Private Mechanisms for Pereference/Vote Data Analyses
Setup experiments in settings_random.py and run main.py
Implemented (almost) all local differential private mechanisms for preference data aggregation
- LAPLACE: add Laplace random noises
- ADDITIVE: use the Additive mechanism
- SAMPLEX0LAPLACE: sampling one candidate then use the Laplace mechanism
- SAMPLEX0BRR: sampling one candidate then use binary randomized response
- SAMPLEX0SUBSET: sampling one candidate then use Subset mechanism
- SAMPLEX1PIECEWISE: use the Piecewise mechanism
- SAMPLEXLAPLACE: optimized weighted-sampling one candidate then use the Laplace mechanism
- SAMPLEXSUBSET: optimized weighted-sampling one candidate then use the Subset mechanism