摘要： 研究水平分布数据集的隐私保护关联规则挖掘算法。针对现有算法需要多次扫描数据集的缺点，提出一种只须对数据集进行2次扫描、基于分布式FP-tree的隐私保护挖掘算法。该算法可以有效降低通信量，能在保证准确度的同时保护原始数据。

关键词: 隐私保护, 分布式关联规则挖掘, 频繁项集, 多方安全计算

Abstract: This paper studies association mining algorithm for privacy preserving on horizontally distributed datasets. Existing algorithms need scan the datasets many times, a new algorithm based on distributed FP-tree is proposed which requires only 2 scans of the datasets. It can lower the traffic effectively and protect the original data while pledging nicety.

Key words: privacy preserving, distributed association rule mining, frequent itemset, secure multi-party computation

中图分类号: 

• TP311