Sensitive attribute privacy preservation of trajectory data publishing based on l-diversity.

Area of Science:

• Computer Science
• Data Privacy
• Geospatial Analysis

Background:

• Ubiquitous positioning technology enables large-scale collection of human movement data for decision-making.
• Raw trajectory data often contains sensitive personal information, posing significant privacy risks upon publication.
• Attacks like record linkage, attribute linkage, and similarity attacks threaten privacy by exploiting background knowledge.

Purpose of the Study:

• To develop a robust data privacy preservation method for trajectory data.
• To mitigate privacy leakage risks associated with spatial-temporal sequences.
• To introduce an enhanced privacy model, (k)-privacy, building upon l-diversity.

Main Methods:

• Identification of critical spatial-temporal sequences prone to privacy leakage.
• Perturbation of these sequences through addition or deletion of spatial-temporal points.
• Ensuring published data adheres to the novel (k)-privacy model, an extension of l-diversity.

Main Results:

• The proposed scheme effectively prevents record linkage, attribute linkage, and similarity attacks.
• Experimental results demonstrate superior privacy preservation compared to existing methods.
• The method maintains high data utility alongside enhanced privacy guarantees.

Conclusions:

• The enhanced l-diversity approach offers a strong solution for privacy preservation in trajectory data.
• The (k)-privacy model provides improved protection against sophisticated privacy attacks.
• This work balances privacy protection with data utility for knowledge discovery from movement data.