摘要
在ECLAT算法的基础上,提出一种新的频繁项集挖掘算法——DS-ECLAT算法。该算法使用回写集和深度搜索最长项集两项新技术,在每次迭代中,无须扫描整个数据库,对于(K+1)项集的探索仅依赖于K项集,并生成K项回写集,下一次迭代时吸取这些回写集,减少了交运算的次数,提高了算法的执行效率。相对于ECLAT算法,新算法减少了内存的需要,具有更好的可伸缩性。
DS(deep search)-ECLAT algorithm is a new frequent itemsets algorithm proposed on the fundation of ECLAT algorithm.In algorithm two new technologies writing-back sets and depth search for the longest itemsets are applied.Through scanning of the whole database becomes evitable in each iterative.Depending only on the K key set,the exploring operation of(k+1) itemsets generates write-back sets of K items,which could be utilized in the next iterative.In this way intersection operations are reduced and algorithm efficiency is improved.Comparing with ECLAT algorithm,DS-ECLAT algorithm requires less memory while acquires better scalability.
出处
《广西科学院学报》
2010年第1期19-22,共4页
Journal of Guangxi Academy of Sciences
基金
广西自然科学基金项目(桂科青0731023)资助
关键词
挖掘算法
频繁项集
回写集
mining algorithm
frequent itemsets
write-back set