Authors:
Feten Ben Fredj
1
;
Nadira Lammari
2
and
Isabelle Comyn-Wattiau
3
Affiliations:
1
CEDRIC-CNAM and Pôle technologique de Sfax, France
;
2
CEDRIC-CNAM, France
;
3
CEDRIC-CNAM and ESSEC Business School, France
Keyword(s):
Anonymization, Privacy, Generalization Technique, K-Anonymity, Algorithm, Guidelines.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Best Practices & Communities of Practice
;
Communities of Practice
;
Computer-Supported Education
;
Information Security
;
Knowledge Management and Information Sharing
;
Knowledge-Based Systems
;
Learning/Teaching Methodologies and Assessment
;
Society, e-Business and e-Government
;
Symbolic Systems
;
Web Information Systems and Technologies
Abstract:
Many techniques, such as generalization algorithms have been proposed to ensure data anonymization before publishing. However, data publishers may feel unable to choose the best algorithm given their specific context. In this position paper, we describe synthetically the main generalization algorithms focusing on their constraints and their advantages. Then we discuss the main criteria that can be used to choose the best algorithm given a context. Two use cases are proposed, illustrating guidelines to help data holders choosing an algorithm. Thus we contribute to knowledge management in the field of anonymization algorithms. The approach can be applied to select an algorithm among other anonymization techniques (micro-aggregation, swapping, etc.) and even first to select a technique.