A type of inference where an agent learns a piece of information about an entity from some data with or without reidentification of that individual in the data. For example, if a dataset tells me that all people living a certain neighbourhood earn less than a certain amount each year and I know that you live in that neighbourhood then I have learnt something about your income.
In principle, the attribution could be deterministic (if X is true then Y is always true as well) or it could be probabilistic (if X is true then Y is more likely to be true than if X is not). The latter is essentially the same as inference and in some parts of the literature a distinction is made between attribute disclosure (deterministic) and inferential disclosure (probabilistic).
Further reading:
See also: ATTRIBUTE, INFERENCE ATTACK
Hittmeir, M., Mayer, R. and Ekelhart, A., 2020. A baseline for attribute disclosure risk in synthetic data. In Proceedings of the Tenth ACM Conference on Data and Application Security and Privacy, 133–43, https://doi.org/10.1145/3374664.3375722.
Rubinstein, I.S. and Hartzog, W., 2016. Anonymization and risk. Washington Law Review, 91, 703–60, https://digitalcommons.law.uw.edu/wlr/vol91/iss2/18/.
Smith, D. and Elliot, M., 2008. A measure of disclosure risk for tables of counts. Transactions on Data Privacy,1(1), 34–52, www.tdp.cat/issues/tdp.a003a08.pdf.
-
Hittmeir, M., Mayer, R. and Ekelhart, A., 2020. A baseline for attribute disclosure risk in synthetic data. In Proceedings of the Tenth ACM Conference on Data and Application Security and Privacy, 133–43, https://doi.org/10.1145/3374664.3375722.
-
Rubinstein, I.S. and Hartzog, W., 2016. Anonymization and risk. Washington Law Review, 91, 703–60, https://digitalcommons.law.uw.edu/wlr/vol91/iss2/18/.
-
Smith, D. and Elliot, M., 2008. A measure of disclosure risk for tables of counts. Transactions on Data Privacy,1(1), 34–52, www.tdp.cat/issues/tdp.a003a08.pdf.
