Class NoisyStats

Object
NoisyStats
All Implemented Interfaces:
Computation<List<DRes<SFixed>>,​ProtocolBuilderNumeric>

public class NoisyStats
extends Object
implements Computation<List<DRes<SFixed>>,​ProtocolBuilderNumeric>
Compute differentially private estimates for the coefficients of a linear model fitted on a dataset. The method is based on the paper Differentially Private Simple Linear Regression by Alabi et. al.
  • Constructor Details

    • NoisyStats

      public NoisyStats​(List<DRes<SFixed>> x, List<DRes<SFixed>> y, double epsilon)
  • Method Details

    • buildComputation

      public DRes<List<DRes<SFixed>>> buildComputation​(ProtocolBuilderNumeric root)
      Specified by:
      buildComputation in interface Computation<List<DRes<SFixed>>,​ProtocolBuilderNumeric>