Skip to main content

Table 4 The lowest MSE for the existing method and the enhancements which were tested on UCI nursery dataset

From: Sarve: synthetic data and local differential privacy for private frequency estimation

Method

ε = ln(2)

ε = ln(3)

ε = ln(4)

ε = ln(5)

ε = ln(6)

ε = ln(7)

RS + FD[ADP]

0.000734962

0.000491035

0.000220841

0.000250292

0.00030747

0.000189085

RAPPOR

0.000981

0.000799

0.000441

0.000529

0.000483

0.000338

Hadamard response

0.00163

0.00075

0.000433

0.000659

0.000556

0.000396

Sarve

0.00087

0.000445

0.000417

0.000259

0.000387

0.000285

Method

ε = 2

ε = 3

ε = 4

ε = 5

ε = 6

ε = 7

RS + FD[ADP]

0.000278

0.000152

0.000111

9.31E−05

8.91E−05

0.000109

RAPPOR

0.000255

0.000219

0.000132

0.000112

8.07E−05

4.76E−05

Hadamard response

0.00048

0.000492

0.000452

0.000376

0.000378

0.000445

Sarve

0.000286

0.000124

0.000112

9.06E−05

0.000102

0.000119

  1. The values in bold indicate privacy conditions when Sarve performed better than adaptive RS+FD and resulted in lower MSE between real and post-privatization estimated frequencies