Scholarlink

FAIRification and open data are relevant practices of heterogeneous data analysis, with an impact on biomedical research transparency. A controversial issue is how to combine FAIR and open data principles for the manipulation of privacy-sensitive health data. Perturbative masking techniques have been used as an alternative for anonymizing numeric and nominal data in health databases to minimize patients’ identity disclosure risks. However, there is still a scarcity of semantic perturbative methods aimed to consider the meaning of data and semantic relationships among concepts. This paper analyzes how perturbative masking methods, enriched with knowledge from ontologies, enable minimizing privacy risks while preserving analytical utility of the anonymized data.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Semantic Perturbative Privacy-Preserving Methods for the Open Release of Health Data

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Mercedes Rodriguez-Garcia,
  • Antonio Balderas,
  • Andres Muñoz,
  • Juan Manuel Dodero

摘要

FAIRification and open data are relevant practices of heterogeneous data analysis, with an impact on biomedical research transparency. A controversial issue is how to combine FAIR and open data principles for the manipulation of privacy-sensitive health data. Perturbative masking techniques have been used as an alternative for anonymizing numeric and nominal data in health databases to minimize patients’ identity disclosure risks. However, there is still a scarcity of semantic perturbative methods aimed to consider the meaning of data and semantic relationships among concepts. This paper analyzes how perturbative masking methods, enriched with knowledge from ontologies, enable minimizing privacy risks while preserving analytical utility of the anonymized data.