- Titre : Investigating different perspectives on vulnerability in mobility datasets
- Présentateur : Lucas Felix (INRIA)
- Résumé : Mobility data offers rich insights into human behavior, enabling applications in transportation, retail, and marketing, among others. However, such data also entails significant privacy risks, as the singularity of individual mobility patterns can reveal user identities. Existing metrics for quantifying user exposure predominantly focus on spatio-temporal singularity (e.g. Uniqueness), overlooking behavioral traits that may also contribute to re-identification. In our ongoing research, we tackle this overlooked part of exposure, evaluating users' behavioral exposure. We propose two novel techniques that quantify the users' exposure. Our contributions are not limited to the quantification, but also the interpretation of the users' exposure, allowing data owners to individually address the exposure of users, potentially maximizing privacy and data utility. We evaluated our proposal in two distinct Call Detail Record (CDR) datasets. Our results suggest that only observing the locations in time of users is not enough to fully capture users' exposure, and that behavior can be a more interpretative proxy for doing so.
- Sujet de recherche associé :