421989.fig.003
Figure 3: Empirical security-utility plot. The plot shows the trade-off between degree of masking and analysis potential for the example data. The points do not fall directly on the curve because the masking methods are not nested (i.e., there is not a one-to-one correspondence between degree of anonymization and analysis potential). The curve is a nonlinear regression fit of the model anonymization score = [ 1 ( a n a l y s i s - p o t e n t i a l s c o r e ) 𝜃 ] 1 / 𝜃 where θ was estimated to be 178.