Privacy homomorphisms for statistical confidentiality

When publishing contingency tables which contain official statistics, a need to preserve statistical confidentiality arises. Statistical disclosure of individual units must be prevented. There is a wide choice of techniques to achieve this anonymization: cell supression, cell perturbation, etc. In t...

Descripción completa

Detalles Bibliográficos
Autor: Domingo Ferrer, Josep
Tipo de recurso: artículo
Fecha de publicación:1996
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2099/4068
Acceso en línea:https://hdl.handle.net/2099/4068
Access Level:acceso abierto
Palabra clave:Computer systems
Statistical confidentiality
Privacy homomorphisms
Ciències de la computació
Classificació AMS::68 Computer science::68P Theory of data
Descripción
Sumario:When publishing contingency tables which contain official statistics, a need to preserve statistical confidentiality arises. Statistical disclosure of individual units must be prevented. There is a wide choice of techniques to achieve this anonymization: cell supression, cell perturbation, etc. In this paper, we tackle the problem of using anonymized data to compute exact statistics; our approach is based on privacy homomorphisms, which are encryption transformations such that the decryption of a function of cyphers is a (possibly different) function of the corresponding clear messages. A new privacy homomorphism is presented and combined with some anonymization techniques, in order for a classified level to retrieve exact statistics from statistics computed on disclosure-protected data at an unclassified level.