Skip to main content

Table 3 Distance to different nearest neighbors compared to the unperturbed data

From: Measuring the impact of spatial perturbations on the relationship between data privacy and validity of descriptive statistics

Perturbation method K = 1 Proportion of original K = 1 K = 5 Proportion of original K = 5 K = 10 Proportion of original K = 10 K = 20 Proportion of original K = 20
Original data 14.57 1 47.67 1 71.42 1 104.09 1
Random perturbation 50 m 20.52 1.41 52.19 1.09 74.28 1.04 106.41 1.02
Random perturbation 250 m 25.07 1.72 62.03 1.3 87.37 1.22 122.83 1.18
Random perturbation 500 m 27.2 1.87 67.12 1.41 95.37 1.34 134.57 1.29
Random weighted perturbation 5 NN 20.38 1.4 48.54 1.02 71.65 1 104.73 1.01
Random weighted perturbation 50 NN 25.02 1.72 60.06 1.26 85.36 1.2 118.45 1.14
Donut between 5 and 50 m 20.5 1.41 52 1.09 74.45 1.04 106.31 1.02
Donut between 50 and 250 m 24.95 1.71 61.28 1.29 86.8 1.22 122.93 1.18
Affine shear 15.2 1.04 48.6 1.02 73.03 1.02 106.97 1.03
Voronoi masked 5.59 0.38 46.38 0.97 70.77 0.99 103.29 0.99
Grid center w/100 m cells 5.94 0.41 34.79 0.73 62.48 0.87 97.84 0.94
Grid center w/250 m cells 1.53 0.11 12.82 0.27 29.1 0.41 62.89 0.6
Grid line w/100 m cells 11.39 0.78 44.58 0.94 69.86 0.98 104.75 1.01
Grid line w/250 m cells 7.31 0.5 32.94 0.69 58.26 0.82 94.25 0.91