Randomization in privacy preserving data mining
ACM SIGKDD Explorations Newsletter
Cardinality-based inference control in data cubes
Journal of Computer Security
| Hi-index | 0.00 |
The increasing demand for information, coupledwith the increasing capability of computer systems,has compelled information providers to reassess theirprocedures for preventing disclosure of confidentialinformation. General logical and numerical methodsexist to determine, prior to release, if disclosurecan occur-either directly or through inference. Onemethod uses linear programming techniques applied tomulti-dimensional tables of count data to determinewhich cells are subject to inferential disclosure. Thispaper develops integer programming techniques (1P)to find an optimal primary suppression set for protectingthe confidentiality of sensitive data in three dimensionaltables. An example is drawn from FederalReserve Bank records. Data tables are randomlygenerated to assess the extent of inferential disclosureand the computational time/space restrictions of theIP model.