Novel Hybrid Geometric Transformation Function to Safeguard LBS Data

In the quest to enhance location data privacy, this investigation embarked on evaluating various geometric transformations' efficacy and this process constructed a custom metric to evaluate safety. Notably, transformations such as RBF, Möbius, Spherical, Tessellation, and Affine were applied to location data sets and then subjected to synthetic adversarial inferences. Our findings show that transformation complexity doesn't guarantee enhanced privacy outright. Specifically, while the RBF and Möbius transformations exhibited high complexity, they were less adept at safeguarding data privacy in certain scenarios. In contrast, simpler transformations on the location data such as Affine surprisingly provided robust protection against synthetic adversaries. The study introduced a new “attack proneness severity” metric, offering insights into each transformation's vulnerability.