Privacy-preserving big data publication: (k, l) anonymity

Abstract

The explosion in variety and volume of information in the public domains provides an enormous opportunity for analysis and business purposes. Availability of private information is of explicit interest in sanctionative highly tailored services tuned to individual desires. Though this is highly favorable to the individuals, the conventional anonymization techniques still possess threats to the privacy of individuals through reidentification attacks. The focus of this paper is to propose a privacy-preserving approach called (K, L) Anonymity that combines k-anonymity and Laplace differential privacy techniques. This coherent model guarantees privacy from linkage attacks as the risk is mitigated through experimental results. The proposed model also addressed the shortcomings of other traditional privacy-preserving mechanisms and validated with publicly available datasets.