Title: Privacy-preservation in data pre-processing for web usage mining

Authors: Michael Mireku Kwakye

Addresses: Department of Computer Science, University of Calgary, 2500 University Drive NW, Calgary, Alberta, Canada; Faculty of Computing and Information Systems, Ghana Technology University College, PMB 100, Accra, Ghana

Abstract: Many online systems offer recommendation of varied aspects to web users to provide enough utility and the right kind of user preference data. Most of these recommendations arise out of well-processed underlying data from data mining procedures. The data mining procedures tend to expose seemingly private data item values to unintended analysts, especially that of demographic data. In this paper, the research study proposes a privacy-preserving framework architecture that enforces privacy on the data item values of web log and demographic data of users. The privacy issues are addressed by introducing a privacy-preserving database that stores a privacy commitment tuple; composed of the data item value and the privacy-preservation property values. This methodology approach proposes to offer merits in ensuring the protection of web users' privacy information during data mining tasks. Additionally, the approach offers different levels of protection; where on the user-centric level, users' personal privacy preferences are enabled, as part of subscribing to the laid down privacy policy on the web system. On the other hand, general privacy protection is enforced at the system-level by service providers (SPs).

Keywords: web usage mining; WUM; data privacy; privacy taxonomy; privacy-preservation; privacy policies.

DOI: 10.1504/IJIPSI.2019.106605

International Journal of Information Privacy, Security and Integrity, 2019 Vol.4 No.2, pp.134 - 152

Received: 07 May 2019
Accepted: 25 Nov 2019
Published online: 15 Apr 2020 *

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