A NOVEL GPU-ACCELERATED APPROACH TO PRIVACY PRESERVATION: ENHANCING MICRO-AGGREGATION WITH ENSEMBLE LEARNING

Abstract

As big data continues to grow in significance across various industries, the challenge of protecting sensitive information becomes increasingly critical. Micro-aggregation is a key technique used to achieve k-anonymity, an essential method for preserving privacy. However, traditional micro-aggregation methods often struggle with scalability and high computational costs when applied to large datasets. This research proposes a novel approach that combines ensemble machine learning techniques (MLT) with GPU (Graphics Processing Unit)-enhanced computation to optimize the micro-aggregation process. By leveraging the parallel processing capabilities of GPUs, our method significantly improves the accuracy of data grouping while reducing computation time. Experimental results demonstrate that the proposed approach enhances privacy protection and maintains data utility, making it well-suited for large-scale data applications. This study provides a scalable and efficient solution for privacy preservation, addressing the limitations of existing micro-aggregation methods.