Enhanced SAC For Text Privacy
TL;DR
An enhanced SAC algorithm in corporating embedding techniques, semantic generalization, and clustering to efficiently process and anonymize text data is introduced to ensure privacy preservation while retaining the utility of the transformed data, making it suitable for various privacy-sensitive applications.
Abstract
Abstract—. Privacy-preserving data publishing has become vital in the age of data-driven decision-making, especially with the increasing availability of unstructured datasets. While the Score, Arrange, and Cluster (SAC) algorithm effectively anonymizes structured data, it does not address the challenges posed by unstructured text data. This proposal introduces an enhanced SAC algorithm in corporating embedding techniques, semantic generalization, and clustering to efficiently process and anonymize text data. The proposed system computes embeddings for textual attributes, creates semantic hierarchies for generalization, and clusters similar records to achieve k-anonymity. This approach ensures privacy preservation while retaining the utility of the transformed data, making it suitable for various privacy-sensitive applications. Keywords- k-anonymity, Generalization, Clustering, Privacy Sensitivity Applications, Semantic Embeddings.