A Differential Privacy-based System for Efficiently Protecting Data Privacy

The daily services people use, such as search engines, mobile services, and online social activity, all have vast sensitive personal information in today’s information realm. The question of how to aggregate sensitive user data without compromising individual privacy is a fundamental obstacle to the increasing data accessibility. Differential privacy has become a widely accepted method for disclosing sensitive information while ensuring statistical privacy. There are numerous algorithms that can be used to address various target functions. However, the research gap identified in the existing works is that the noise-added data used in the algorithm is in decimal values, and in some cases, it is negative as well. In general, count or frequency can neither be negative nor decimal, but should rather be a whole number. This research focuses on developing a unique mechanism to protect the sensitive information with the use of differential privacy. Laplace noise addition is frequently advanced as a method for satisfying differential privacy. A dataset is curated by collecting user responses by a survey form and stored as an actual dataset to which the random noise is added to produce the noise added dataset using the differential privacy algorithm i.e., Laplace algorithm. The randomized response algorithm is used to create the noise added dataset. Laplace Mechanism Algorithm is implemented by C++ and Rust, and the result is compared with the existing algorithms. The algorithm only produces whole number that is integers as the count values and produced results with better accuracy than the existing Laplace algorithm.