Protecting privacy and enhancing utility: A novel approach for personalized trajectory data publishing using noisy prefix tree

Abstract

In recent years, the widespread adoption of location-based software has significantly improved people’s daily lives. However, this convenience has brought about an increasingly severe concern: the risk of data privacy breaches. To tackle this issue, a novel scheme for personalized trajectory data publishing is proposed, leveraging a noisy prefix tree structure. The scheme begins by constructing multiple trajectory equivalence classes based on the spatiotemporal characteristics of trajectories, followed by the calculation of distinct Hilbert curve orders for each equivalence class. Subsequently, the Hilbert curve is employed to partition the location points within each equivalence class, with the aid of a scoring function that selects optimal center points to replace other location points effectively. Additionally, appropriate encoding levels are determined for each partitioned region, facilitating the conversion of the center points into binary encoded characters using Hilbert-Geohash. This conversion process safeguards against data privacy leakage by obfuscating the original latitude and longitude coordinates. Ultimately, a noisy prefix tree is constructed to store the binary encoding in its nodes. To ensure data privacy preservation, a novel privacy budget allocation approach is introduced, applying suitable Laplace noise to each node. The effectiveness of the proposed algorithm is confirmed through experimental comparisons with existing schemes, employing real datasets and demonstrating its notable performance in terms of both data privacy and utility.