A Privacy Index Calculator for Federated Medical Studies

Sharing clinical datasets requires rigorous privacy guarantees to prevent subject re-identification. This paper presents a privacy analysis tool that operates entirely client-side, ensuring data protection by eliminating the need for server-side processing. The system implements a hybrid classification mechanism that combines semantic rules with statistical heuristics to automatically categorize attributes. Furthermore, a modular, plugin-based architecture computes a comprehensive Privacy Index by aggregating multiple privacy models and detecting existing protection techniques through pattern recognition. Validation on synthetic datasets demonstrated the tool's effectiveness in correctly identifying high-risk attributes and accurately quantifying privacy guarantees.