Scholarlink

<p>In the realm of big data analytics, safeguarding individual privacy emerges as a critical challenge. While data anonymization models serve to address this concern, they often prioritize privacy at the expense of data utility. This paper investigates an augmented multidimensional bottom-up anonymization methodology within the Apache Spark environment, tailored to adhere to the k-concealment privacy paradigm. To the best of our knowledge, this is the first study to employ the k-concealment model in the context of big data publishing. Our approach hinges on refined appending and partitioning procedures. The former entails the aggregation of data records from diverse sources into an enhanced bottom-up R-tree generalization, minimizing information loss. Meanwhile, the latter focuses on partitioning overflowed nodes in accordance with the k-concealment criteria, thereby mitigating overlap between resultant partitions. Empirical findings underscore significant enhancements in both data utility and processing time compared to prevalent top-down approaches documented in existing literature.</p>

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

BuKc: A novel bottom-up approach for enhanced data anonymization in apache spark

  • 期刊论文

  • 发表日期：

    2026年1月3日

查看全文
  • Salheddine Kabou,
  • Laid gasmi,
  • Abdelbaset Kabou

摘要

In the realm of big data analytics, safeguarding individual privacy emerges as a critical challenge. While data anonymization models serve to address this concern, they often prioritize privacy at the expense of data utility. This paper investigates an augmented multidimensional bottom-up anonymization methodology within the Apache Spark environment, tailored to adhere to the k-concealment privacy paradigm. To the best of our knowledge, this is the first study to employ the k-concealment model in the context of big data publishing. Our approach hinges on refined appending and partitioning procedures. The former entails the aggregation of data records from diverse sources into an enhanced bottom-up R-tree generalization, minimizing information loss. Meanwhile, the latter focuses on partitioning overflowed nodes in accordance with the k-concealment criteria, thereby mitigating overlap between resultant partitions. Empirical findings underscore significant enhancements in both data utility and processing time compared to prevalent top-down approaches documented in existing literature.