A New $k$k-Anonymity Method Based on Generalization First $k$k-Member Clustering for Healthcare Data

Advances in microelectronics and the evolution of IoHT devices allow the collection and, consequently, generation of a greater volume of health data, intensifying the need for robust data privacy solutions. Traditional <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">k</i>-anonymity-based anonymization techniques often suffer from high information loss, especially as the anonymity parameter k increases. To address these challenges, this article proposes Generalization First <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">k</i>-Member Clustering (GFKMC), a novel <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">k</i>-anonymity method that applies early generalization to quasi-identifiers, reducing computational overhead and minimizing information loss. Unlike traditional methods (e.g., Mondrian, Top-Down Greedy (TDG), and Clustering-Based (CB)), GFKMC maintains nearly constant information loss (≈ 25%) across varying <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">k</i> values and better preserves machine learning model performance, especially in low <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">k</i> scenarios. Empirical evaluations demonstrate that GFKMC outperforms baseline methods by significantly minimizing the trade-off between data utility and privacy. Moreover, GFKMC preserves the performance of machine learning models more effectively.