Three-way optimization of privacy and utility of location data

With the recent bloom of data and the drive towards an information-based society, the urge of and the advancements in data analytics is surging like never before. And with this, the risks of privacy violation of various kinds are also increasing manifold. Most of the methods to mitigate the privacy risks for location data resort to adding some noise to the location, like the planar Laplace used to achieve geo-indistinguishability. However, the noise should be calibrated carefully, taking into account the implications for utility, because it is far from ideal for the service providers to completely lose the utility of the collected data succumbing to the privacy requirements of the users. Similarly, the quality of service for the users should be optimized with their personalized needs of privacy protection used to shield their sensitive information. In this paper, we address this age-old battle between privacy and utility from three ends: privacy of the users' data, the quality of service (QoS) received by them in exchange for sharing their privatized data, and the statistical utility of the privatized data for the service providers who wish to perform various kinds of analysis and research on the data collected from the users. We propose a method to produce a geo-indistinguishable location-privacy mechanism that advances to optimize simultaneously between the level of privacy attained, the QoS, and the statistical utility achieved by the obfuscated data. We illustrate the soundness of this three-way privacy-utility optimization mechanism both analytically and with experiments. Apart from the novelty of the proposed method, this work is aimed to engender an analytical perspective to bridge between geo-indistinguishable location-privacy, QoS, and statistical utilities used in standard data analytics, from an information theoretical, probabilistic, and statistical perspective.