基于双向十字链表的频繁项集挖掘

有效地进行频繁项挖掘一直以来都是数据挖掘任务中最为重要的组成部分。已有的大部分频繁项挖掘算法在数据项多及支持度低的情况下,算法的效率急剧下降。为了有效地解决此类问题,提出了一种采用双向十字链表结构的频繁项挖掘算法(two-way crossed list for frequent itemsets mining,TCLFI)。极大地降低了搜索空间,加快了频繁项的筛选过程,减少了所需保存的数据项个数,从而降低了时间复杂度,提高了频繁项的挖掘效率。实验通过真实数据集和合成数据集验证了算法的有效性和扩展性。

Frequent Itemsets Mining Based on Two-way Crossed List

Frequent itemsets mining is one of the key processes of Data mining. The time complexity is very high, especially in the case of large itemset and low minimum support, in existing algorithms. This paper presents a frequent itemset mining algorithm named TCLFI (two-way crossed-linked list for frequent itemsets mining) to improve mining efficiency. It can decrease the search space and the scale of itemset, and also can speed up the selection of frequent itemset with mapping transaction data into a two-way cross-linked list structure to mining the frequent pattern. Experiments with real and synthetic datasets validated the effectiveness and scalability.