基于关联规则的模式库划分算法

AN ALGORITHM FOR PATTERN BASES PARTITION BASED ON ASSOCIATION RULES

在多数据库挖掘的过程中一般要先将多个数据库按照某种规则进行划分,再进一步进行模式挖掘,提出了一种基于关联规则的相似度测量方法,将各个局部模式库进行划分,并对划分的结果进行评价,接着根据评价的结果设计出了一个选择最好划分的算法,找出最好的一种划分。最后经实验验证,算法是准确而有效的。

Multi-databases are partitioned according to some rule before pattern mining. A similarity measuring method based on association rules is proposed. The local pattern bases are separated, and the results are evaluated. An algorithm for choosing the best partition result is designed based on the evaluation results, and the best division is found out. The experimental results show that the proposed algorithms are accurate and effective.