Contributions to Statistical Disclosure Control: Enhancing Multivariate Microaggregation using Graph Theory

Identifier:  TDX:4146

Handle: https://hdl.handle.net/20.500.11797/TDX4146

Authors:  Maya López, Armando

Abstract:
This PhD dissertation focuses on microaggregation as a technique for statistical disclosure control. The research aims to enhance the efficiency and quality of fixed-size and variable-size microaggregation through the use of the Travelling Salesman Problem (TSP) algorithm. The study presents three contributions: (1) TSP-based fixed-size microaggregation, (2) TSP-based variable-size microaggregation, and (3) dataset reduction strategies for TSP-based microaggregation. The proposed methods are evaluated through experiments and compared with existing techniques. The research findings reveal that the proposed TSP-based methods outperform the existing ones in terms of both data utility and computation time. This dissertation provides a comprehensive study of the state-of-the-art in statistical disclosure control, and offers practical solutions to enhance the performance of microaggregation for privacy-preserving data publishing.