Skip to main content

Table 3 Estimated coefficients of the logistic regression model predicting survival of Titanic passengers using the original (left) and the anonymised (right) data

From: A deterministic approach for protecting privacy in sensitive personal data

Coefficient Original data Anonymised data Std. difference
Estimate Std. Err. z value Pr(\(>|z|\)) Estimate Std. Err. z value Pr(\(>|z|\))
Intercept 3.519 0.437 8.054 8.00e–16 3.615 0.454 7.972 1.56e–15 0.220
Pclass (2nd) − 1.067 0.286 − 3.732 0.0002 − 1.112 0.300 − 3.707 0.0002 0.159
Pclass (3rd) − 2.283 0.281 − 8.135 4.12e–16 − 2.343 0.300 − 7.822 5.21e–15 0.216
Sex (male) − 2.628 0.194 − 13.527 < 2e–16 − 2.625 0.194 − 13.508 < 2e–16 0.012
Age − 0.033 0.008 − 4.437 9.13e–06 − 0.035 0.008 − 4.586 4.53e–06 0.205
Fare 0.001 0.002 0.466 0.641 0.001 0.002 0.217 0.829 0.223
Family (yes) − 0.091 0.194 − 0.471 0.638 − 0.089 0.197 − 0.452 0.651 0.010