关联规则算法优化研究与实现

本文主要针对Apriori算法采用最小支持度和最小信任度阈值来发现知识,而没有考虑交易中数量问题的不足,提出一种快速的基于频繁模式树FP-tree的最大频繁项目集挖掘算法.该算法不需要产生频繁项集,而且只需要扫描事务数据库D一次,从而提高了算法的执行效率.该方法结合大量的实际项目数据进行关联规则挖掘测试发现,不仅能较好地分析非稠密数据,也能处理现实世界中稠密数据. 结果 表明该优化算法可显著降低关联规则挖掘在数据挖掘工作中的时间开销.

Research and Implementation of the Optimized Assocation Rules Mining Arithmetic

In this paper, Apriori algorithm for minimum support and minimum confidence threshold to discover knowledge, without taking into account the number of transactions of less than, a fast tree-based fi＇cquent pattern FP-tree of the maximum frequent itemsets mining algorithm. The algorithm does not require frequent itemsets generated, and only need to scan a database D, thereby enhancing the efficiency of the algo- rithm. The combination of a large number of actual project data mining association rules found not only a better analysis of non-dense data, but also be able to handle real world data populated. The results show that the optimization algorithm can significantly reduce the associated regu- lation of data mining excavation work in the time overhead.