Scholarlink

<p>The multi-party Privacy-Preserving Record Linkage (PPRL) aims to identify and match the same entity across different parties’ data sources while ensuring that all private data remains protected and undisclosed, except for the final matching results shared among the parties. The previously proposed a multi-party PPRL method based on secondary encoding suffers from issues related to the number of data splits and load balancing, which impact computational efficiency and linkage quality. We propose an extended approach—a multi-party PPRL method based on improved secondary encoding(ISE_PPRL). By adopting a rational data split strategy and participant work strategy, and incorporating the geometric mean and consistent hashing algorithm, this method overcomes the traditional limitation where each party could only process a single data split. It effectively mitigates potential collusion risks among data splits, enhances security, optimizes computational efficiency, and maintains linkage quality. Experimental results demonstrate that this method exhibits significant advantages in improving security, enhancing computational efficiency, and maintaining linkage quality.</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

A multi-party privacy-preserving record linkage method based on improved secondary encoding

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年7月2日

查看全文
  • Shumin Han,
  • Yizi Wang,
  • Derong Shen,
  • Chuang Wang

摘要

The multi-party Privacy-Preserving Record Linkage (PPRL) aims to identify and match the same entity across different parties’ data sources while ensuring that all private data remains protected and undisclosed, except for the final matching results shared among the parties. The previously proposed a multi-party PPRL method based on secondary encoding suffers from issues related to the number of data splits and load balancing, which impact computational efficiency and linkage quality. We propose an extended approach—a multi-party PPRL method based on improved secondary encoding(ISE_PPRL). By adopting a rational data split strategy and participant work strategy, and incorporating the geometric mean and consistent hashing algorithm, this method overcomes the traditional limitation where each party could only process a single data split. It effectively mitigates potential collusion risks among data splits, enhances security, optimizes computational efficiency, and maintains linkage quality. Experimental results demonstrate that this method exhibits significant advantages in improving security, enhancing computational efficiency, and maintaining linkage quality.