Machine-Learning-Based Re-identification of Location Histories Using Features Considering Sparsity Obtained from Social Networking Services | IEEE Conference Publication | IEEE Xplore

Machine-Learning-Based Re-identification of Location Histories Using Features Considering Sparsity Obtained from Social Networking Services


Abstract:

Location histories collected from smartphone GPS and LTE base stations are being used by more and more services. Even if they are anonymized to protect personal informati...Show More

Abstract:

Location histories collected from smartphone GPS and LTE base stations are being used by more and more services. Even if they are anonymized to protect personal information, a re-identification attack can be used to de-anonymize the information. The method for de-anonymizing location histories devised by Ohka et al. does not consider the sparsity of location history. We have improved their method by adding two features: compressed visit information and displacement time ratio. Experiments demonstrated that adding these features improves the re-identification rate.
Date of Conference: 29 October 2024 - 01 November 2024
Date Added to IEEE Xplore: 28 November 2024
ISBN Information:

ISSN Information:

Conference Location: Kitakyushu, Japan

Contact IEEE to Subscribe

References

References is not available for this document.