基于改进倒排表和集合的最频繁项集挖掘算法

摘要

最频繁项集挖掘是文本关联规则挖掘中研究的重点和难点,它决定了文本关联规则挖掘算法的性能.针对当前在最频繁项集挖掘方面的不足,将集合论引入倒排表以对其进行改进,然后以此为基础提出了几个命题和推论,并结合最小支持度阈值动态调整策略,提出了一个基于改进的倒排表和集合理论的最频繁项集挖掘算法,最后对所提算法进行验证.实验结果表明,所提算法的规则有效率和时间性能比常用的两个最频繁项集挖掘算法,即NApriori和IntvMatrix算法都好.%Most frequent item sets mining is the focus and the difficulty of text association rules mining, and directly determines the performance of text association rules mining algorithms. Aiming at shortcomings existing in most frequent item sets mining algorithms, this paper improved traditional inverted list, it combined minimum support threshold dynamic adjustment strategy and presented a new most frequent itemset mining algorithm based on improved inverted list and set theory. In addition , it also offered several propositions and deductions which were used to improve the performance of the provided algorithm. Finally, through experiment testing, the provided algorithm is better in effective rate of rules and time performance than NApri-ori and IntvMatrix which are two frequently-used most frequent itemsets mining algorithms.