Interval Privacy: A Framework for Privacy-Preserving Data Collection

The emerging public awareness and government regulations of data privacy motivate new paradigms collecting analyzing that are transparent acceptable to owners. We present a concept corresponding formats, mechanisms, theories for privatizing during collection. privacy, named Interval Privacy, enforces the raw conditional distribution on privatized be same as its unconditional over nontrivial support set. Correspondingly, proposed mechanism will record each value random interval (or, more generally, range) containing it. mechanisms can easily deployed through survey-based collection interfaces, e.g., by asking respondent whether is within randomly generated range. Another unique feature they obfuscate truth but do not perturb Using narrowed range convey information complementary popular paradigm perturbing data. Also, generate progressively refined at discretion individuals, naturally leading privacy-adaptive develop different aspects theory such composition, robustness, estimation, regression learning from interval-valued provides perspective human-centric where individuals have perceptible, transparent, simple way sharing sensitive