zhao

@yongxin zhao

7사용 방법

0공유하기

0좋아요

0Fish Audio에 의해 저장됨

We measure data utility through query accuracy—Mean Absolute Error and Mean Relative Error on trajectory count queries—and spatial fidelity using Hausdorff distance. We also measure computational efficiency through construction time.Our results show that PerTrajTree-DP consistently achieves the lowest query error across all privacy budgets. At ε_total = 1.0, PerTrajTree-DP achieves a Mean Absolute Error of approximately 5.8, significantly outperforming AHGH/PT with 8.2, DP_SR with 12.5, and QJLP with 15.1—about a 40% improvement over the best structural baseline. The Mean Relative Error follows a similar trend.