分布式数据库的关联规则更新算法

Updating algorithm of association rules for distributed database

提出一种分布式数据库的频繁项目集更新算法.该算法主要考虑分布式数据库记录总数不变,增加新项目集后的频繁项目集更新情况.算法排除原数据库已挖掘的频繁项目集,减少了各站点候选频繁项目集数目,同时减少了各站点之间传送的频繁项目集数目,减少网络流量,提高了频繁项目集挖掘的效率.通过理论分析,该算法比FDM算法效率高,并通过实例和实验证明了算法的有效性和可行性.

In this paper, we propose an updating algorithm of association rules for distributed database on condition that the length of transaction database DB unchanged and itemsets I changed. The algorithm removes the frequent items which have been gotten in original database DB to reduce the candidate frequent items and frequent items to be sent between each site and to reduce network current capacity. The efficience of the algorithms has been greatly improved. Through the theoretical analysis, the efficiency of the algorithm is higher than the FDM algorithm. The example analysis and the experiment shows the validity and the feasibility of the algorithm.