A similar paper, ~3k citations, on deanonymization in sparse datasets: https://systems.cs.columbia.edu/private-systems-class/papers/Narayanan2008Robust.pdf htt

A similar paper, ~3k citations, on deanonymization in sparse datasets: https://systems.cs.columbia.edu/private-systems-class/papers/Narayanan2008Robust.pdf https://systems.cs.columbia.edu/private-systems-class/papers... It's well known within the sphere that anonymization is hard and deanonymization is trivial. The original Netflix prize around their recommender system also had issues with deanonymization.