Title: Heuristic-based hybrid privacy-preserving data stream mining approach using SD-perturbation and multi-iterative k-anonymisation

Authors: Paresh Solanki; Sanjay Garg; Hitesh Chhinkaniwala

Addresses: Department of Computer Science and Engineering, Nirma Institute of Technology, Nirma University, Ahmedabad, Gujarat, India ' Department of Computer Science and Engineering, Nirma Institute of Technology, Nirma University, Ahmedabad, Gujarat, India ' Adani Institute of Infrastructure Engineering, Ahmedabad, Gujarat 382421, India

Abstract: Different e-sources are regularly generating huge volumes of data. Data mining is the technique of gathering knowledge from a dataset, but dataset often contains sensitive information so discharging such data may cause privacy breaches. The problem of privateness desires to be is addressed earlier than streaming facts are launched for mining and evaluation functions. Various algorithms proposed so far have focused mainly on static data and very few are on data streams. Perturbation and k-anonymity have received significant attention over other privacy-preserving techniques because of its easiness and effectiveness in guarding data. The proposed hybrid approach is an extension to heuristic-based data perturbation where privacy is preserved through computed tuple values for each instance and users define sensitive drift (SD) and an extension to k-anonymisation where privacy gain has been worked out for choosy anonymisation for a set of tuples and perturbs the sensitive attribute values on data streams.

Keywords: stream mining; anonymisation; perturbation; privacy-preserving; data mining.

DOI: 10.1504/IJKEDM.2018.095522

International Journal of Knowledge Engineering and Data Mining, 2018 Vol.5 No.4, pp.306 - 332

Received: 23 Apr 2018
Accepted: 10 Jul 2018
Published online: 08 Oct 2018 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article