基于二分搜索的最大频繁项目集求解算法改进  

IMPROVING SOLUTION ALGORITHM OF MAXIMUM FREQUENT ITEMSETS BASED ON BINARY SEARCH

作　　者：郑全朝[1] 武建华[2] 

机构地区：[1]暨南大学信息科学技术学院,广东广州510632 [2]暨南大学珠海学院,广东珠海519070

出　　处：《计算机应用与软件》2010年第5期269-271,277,共4页Computer Applications and Software

摘　　要：针对Apriori算法的两个性能瓶颈以及New_Apriori算法中不可忽略的问题,提出新的算法TBapriori。该算法通过二分搜索方法改变由低维到高维或者从高维到低维的搜索策略,直接从1-频繁项目集查找最大频繁项目集,减少了扫描数据库的次数及生成候选集的数量,从而减少发现最大频繁项目集的时间,实验结果表明该算法提高了运算效率。In this paper a new algorithm called TBapriori is presented to overcome two bottlenecks in Apriori algorithm and the non-negligible problem in New_Apriori algorithm.Making change with binary search method on previous search strategy which searches from low-dimension to high-dimension or reversely,in new algorithm it directly searches the maximal frequent itemsets from 1-frequent itemsets and cuts down the times of scanning the database and the number of candidate sets generated,therefore reduces the time of finding maximal frequent itemsets.Experiments results show that the algorithm improves the operation efficiency.

关 键 词：关联规则 APRIORI算法 二分搜索 最大频繁项目集 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]