Towards Quantifying Privacy in Process Mining

TL;DR

This paper introduces a method to quantify privacy preservation effectiveness in process mining by measuring disclosure risks and data utility, validated on real-life event logs.

Contribution

It proposes novel quantitative measures for assessing privacy protection and data utility in process mining, addressing a gap in evaluation methods.

Findings

Effective measures for disclosure risk evaluation

Quantitative assessment of data utility preservation

Validation on multiple real-life event logs

Abstract

Process mining employs event logs to provide insights into the actual processes. Event logs are recorded by information systems and contain valuable information helping organizations to improve their processes. However, these data also include highly sensitive private information which is a major concern when applying process mining. Therefore, privacy preservation in process mining is growing in importance, and new techniques are being introduced. The effectiveness of the proposed privacy preservation techniques needs to be evaluated. It is important to measure both sensitive data protection and data utility preservation. In this paper, we propose an approach to quantify the effectiveness of privacy preservation techniques. We introduce two measures for quantifying disclosure risks to evaluate the sensitive data protection aspect. Moreover, a measure is proposed to quantify data utility preservation for the main process mining activities. The proposed measures have been tested using various real-life event logs.