An approach to custom privacy policy violation detection problems using big social provenance data

Abstract

Social media software changes its system-wide privacy policies over time. Changes in such system-wide policies affect the privacy policies of the individual users. In turn, social media users lack the ability to configure and enforce customized privacy policies according to the precise privacy measure that they demand. To this end, we argue that there is an emerging need for third-party solutions that are independent of existing social media software and that can detect custom privacy policy violations. In order to address this need, we propose an approach to detecting custom privacy violations for social media users. We also introduce a generic software architecture that can be integrated with existing social media software to enable users to keep track of their data. The proposed solution utilizes social provenance data, which is defined as the metadata that describe the lifecycle of the data. To facilitate testing of the software architecture, we developed a prototype implementation and generated a large-scale synthetic provenance dataset. We discuss the details of the prototype implementation and the synthetic dataset. We evaluate the performance of the prototype under an increasing workload. We show the usability of the proposed architecture, as the initial performance testing results are promising.