基于模式分解树的增量挖掘

数据挖掘具有广泛的应用,频繁模式发现是关联规则挖掘问题的重要组成部分。频繁模式的增量挖掘是一个挑战性的任务,已有的几种基于Apriori思想的方法,具有代价太高的弱点。本文提出了一个基于模式分解树,不需要扫描原数据库的增量挖掘算法。通过合理地组织候选项索引,可以取得较高的效率。

Incremental Mining Based on Pattern Decomposing Tree

Data mining technique has its extensive applications. Frequent pattern finding is an important component of association rules mining problem. Incremental mining for frequent pattern is a challenging task. Some existing approaches which are all based on Apriori are too expensive .The paper presents an algorithm for incremental mining which is based on pattern decomposing tree and has no need for scanning old data base. By carefully organizing candidate item set, it can get good efficiency.