Enhancing Privacy in Location-based Services using k-Anonymity, l-Diversity, and Differential Privacy
Selvarajan Saraswathi · 2025
Introduction Location-Based Services (LBS) are an amenity that has taken over modern applications, offering direction, location sharing, and a set of personalized recommendations in real-time. But there are much concerns for privacy in many applications of LBS. The location of users can be used for tracking, profiling and unauthorized access. In light of those challenges, this work introduces an advanced privacy-preserving architecture in LBS, which incorporates k-anonymity, l-diversity, along with differential privacy for improving user privacy. We address the above issues by employing a k-anonymity approach: thus to make a user’s location indistinguishable from at least k other users (k-anonymity), and preventing attribute disclosure of a group by maintaining at least l - diverse sensitive (i.e., location) values within a group (l-diversity). Differential privacy, by contrast, adds controlled noise to the location data, a process that protects location data from being traced back to an individual but allows the data to still be used. By combining techniques in transfer learning and inherent privacy reductions, we hope to prevent common inference attacks and prevent re-identification common with other anonymization techniques. To verify the effectiveness of our framework, we perform extensive simulations based on real-world LBS datasets and compare the performance of our protection mechanism with existing privacy-preserving solutions. Experimental results validate that our approach has a good trade-off between protecting location privacy and utility of the data, by effectively diminishing location privacy threats and having high accuracy of services. We further study computation efficiency and privacy trade-offs and demonstrate the flexibility of our framework for different LBS contexts, such as navigation services, geotagging and smart city applications. The results show that our approach facilitates the development of privacy enhancing techniques for LBS services. Adaptive mechanisms may provide more precise privacy permissions, which future work can integrate.