Personalized track data privacy protection method based on semantics

Abstract

The invention provides a personalized track data privacy protection method based on semantics. According to the method, a generalization region comprising a plurality of similar interested points is respectively constructed according to each sensitive position point provided by users, one similar interested point is selected for replacing the sensitive position point, and in addition, a reasonable replacing track path and a replacing position point sequence are constructed in the generalization region. The personalized track data privacy protection method has the advantages that the requirement of personalized track data privacy protection of the users is met, and meanwhile, the replacing track path data can be enabled to have higher effectiveness.

Description

technical field [0001] The invention relates to data publishing in the communication field, in particular to a semantic-based privacy protection method for individualized trajectory data. Background technique [0002] In recent years, with the widespread application of location-aware devices such as RFID, GPS, and smartphones, more and more location data are collected, which are stored in databases in the form of trajectories. Trajectory data contains a wealth of knowledge, and publishing these data for mining and research by relevant departments can support a variety of related applications, such as traffic planning, location-based advertising, and wildlife tracking. However, these trajectory data often contain private data related to personal sensitive information. If these trajectory data are not processed and released directly, it will cause serious privacy leakage. The privacy leakage in the trajectory data release process can be roughly divided into two categories: on...

AI Technical Summary

Problems solved by technology

The Chinese invention patent with the publication number CN20141008673 discloses a system and method for privacy protection of frequency-based trajectory suppression data release, which uses a specific trajectory local suppression method for anonymity processing, and first finds the privacy data that does not satisfy the user from the original trajectory data set. The problematic projection set of tolerance, then sort the problematic projection set by the frequency of occurrence in the original trajectory data set, and save the result to the new data set, next, find the smallest violation of privacy in the new data set The required trajectory sequence set removes the data that has reached a certain set value in the trajectory sequence set and the user's privacy correlation and data utility. Although the frequency problem is considered to divide the sensitivity of the location point, and the location point with a high sensitivity is suppressed, Improve the utility of anonymized data, but the suppression method has been proved not enough to protect the privacy of mobile users, the attacker can reconfirm the identity of the user who belongs to the trajectory by associating external knowledge, and then obtain the user's private information through the trajectory

In order to solve this association attack, trajectory k-anonymity based on the idea of ​​generalization is proposed, and the trajectory satisfying k-anonymity is put in the same anonymous set. However, when the data comes from transaction records, RFID data and purchase records, The utility of k-anonymized data remains to be tested

In addition to the suppression method and generalization method, the false data method is also a commonly used trajectory data release privacy protection method. The false data method artificially constructs a certain number of false trajectories to reduce the risk of disclosure of real trajectories. However, the current false data methods mainly Replace the original sensitive points with the position points around the sensitive points in the trajectory data, the data utility is not high

Images