增量式的多源序列模式挖掘隐私保护算法

Efficient incremental algorithm for mining sequence patterns in multiple databases with privacy preserving

为解决多数据源挖掘隐私保护问题,文章采取按相似度分类多源数据库及其增量数据库,利用原始数据库挖掘结果和增量数据库分析结果进行敏感序列模式匹配,以有效减少数据库扫描次数的方法,设计实现隐私保护的增量式的高投票率序列模式挖掘算法。实验结果表明,给出的算法既能够准确挖掘出多数据源中全局高投票率模式,又能有效地隐藏保护敏感模式,且显著缩短了挖掘时间。

To solve the problem for incremental mining sequence patterns in multiple databases with privacy preserving, the original databases and their corresponding incremental databases are classified by similarity of item set, the sensitive sequential pattern set is used to match each class of original databases and the incremental databases to reduce the scan times, and a privacy-preserving incremental algorithm for mining the global high-voting sequential patterns in multiple databases is designed. The experimental results show that the proposed algorithm can mine correctly the global high-voting sequential patterns in multiple databases, hide the sensitive patterns, and shorten remarkably the mining time.