BM (Break-Merge): An Elegant Approach for Privacy Preserving Data Publishing

Publishing of person specific data has elevated much concern on the individual privacy. Many frameworks and privacy principles were proposed to protect the privacy of the published data. However, techniques must be investigated onattacker's background knowledge. This paper proposes a new approach, Break-Merge (BM) to reduce the associations between quasi-identifiers and sensitive attributes in an anonymized data. On the fly our approach reduces the attacker's inferring nature on sensitive data drastically by decomposing the anonymizedtable into Quasi-identifier table (QIT) and Sensitive attribute tables (ST's).